CREDIT SUISSE esearch Analyte suaaegacnascet see com 5 staramangoreat suse cr cncommunnengioedtsure cr tongue suse cr vas ono susse com vonnagowsteases.cn sehaleutaagioud austen Inomoson wteet-ten com 16 March 2011 ‘AmericasiUnited States Equity Research WT Hardware / OVERWEIGHT IT Hardware Serving the Smart Era = From Cyclical to Structural Growth. Based on an extensive study of end demand for the IT hardware and IT services sector, we see the $'.21n market growing 4% per annum trough 2015. Our analysis suggests that IT is being Under consumed by the global economy when We look at levels of net tech investment, especialy gven the healthy corporate backdrop. Such Lnderinvestment comes at a tine when we believe IT infrastructure needs to cope with the megatrends of vitualzation, move toward cloud computing, and explosive growth n smart devices, including smartphones and tablets. Given the above and sector valuations, we inate coverage ofthe sector at Overweight = PCs forecasting: Think Compute; Tablets a $1206n Opportunity. We have developed what we believe isthe first econometric model for projecting computing demand. When combined with our commercial PC model, we conclude that installed base for computing products will rise to 2.4bn by 2015, driven by a high elasticity of demand and resulting in 14% unit and 4% revenue growth long term. Our proprietary PC price tier analysis suggests tablet unit volumes of 65mn/1 18mn in 2011/12 and long-term revenue of $120bn which represents 42%4(42% of overall PC industry volumesivalue. = Smartphones, Storage, and Services Driving the Market, while Traditional PCs and Printing and Servers Face Headwinds. Beyond tablets, we see healthy 14% revenue growth within smartphones (we are ‘again raising our estimates), 10% revenue growth in networked storage, and 5% revenue growth in services. By contrast, we expect the PC industry ‘excluding tablets to dectine 6%, printing (hardware and supplies) to remain flat, and servers to decline 2% long term (through 2015), = Apple, EMC, and HP Aro Outperform Rated, while IBM, Xerox, and NetApp Are Neutral Rated; Dell and Lexmark Are Underperform Rated. ‘Apple is our top pick, and we inate with EPS estimates that are 10%/24% above consensus for FY11/12 and see upside to $500. We believe Apple can maintain its compettive advantage in the industry owing to a well ‘esigned vertically integrated model and drive near 50% top and bottom-line growth long term driven by iPhone and iPad. EMC is a direct play on networked storage and virtualzation and given the scope for sustained share gains, bottom-line growth should be robust at 20% per year, meaning ENC should trace at our target ex-cash PIE of 16x, giving upside to $34. For Dell, we initiate with an Underperform rating owing to anemic top-line growth ‘and only gradusl expansion in margins. Furthermore, any growth through MAA is risky, in our view as rivals have deeper pockets and the company’s, track record is mixed. For Lexmark, while valuation is not demanding, we expect negative EPS momentum through 2011 to drive underperformance as the cyclical talwinds that the company enjoyed in 2010 fade. Credit Su DISCLOSURE APPENDIX CONTAINS IMPORTANT DISCLOSURES, ANALYST CERTIFICATIONS, INFORMATION ON TRADE ALERTS, ANALYST MODEL PORTFOLIOS AND THE STATUS OF NON-U.S ANALYSTS. U.S. Disclosure does and teeks to do busine ‘aware that the Fim may have a confit of interest that could affect the objectivly of this report. Investors should consider tis ‘eport as only a sing facto’ in making heir investment docsion. ‘with companies covered in ts esearch reports. As a resull, investors should beCarprrsuss™ Table of contents Serving the Smar Ten Themes for the Next Five Years 1) Gyclical Tailwinds for Now, Structurally 4% Growth 2) Compute Demand Drives Tablets to $1206n Market 43) Smartphones Set to Cross thn Unit Mark by 2015 4) Storage Has Several Secular Growth Drivers 65) Services Set for a Gradual Recovery 6) Lackluster Outlook—Servers, Traditional PCs, Printing 7) Radical Changes to the PC Value Chain? 8) Competition for the Datacenter... Heating Up 9) M&A—Sector Is Ripe for Continued Consolidation 10) The Cloud May Not Be That Incremental Macro Trends Supportive of Tech Get Your Head in the Clouds PGs Disruption Coming PC—Winners and Losers ‘Smartphone Growth to Continue ‘Smartshones—Winners and Losers IT Services—2011 Shows a Gradual Recovery Storage—Secular Growth Storage ~ Winners and Losers Servers—Anemic Growth Ahead Servers ~ Winners and Losers Printing—Secular Challenges 2 18 4 “ 16 15 18 23 25 28 37 R sor 128 181 197 235 258 281 305 324Carprrsuss™ Serving the Smart Era Over the past three years, the IT sector has grown at a 1% CAGR. On the enterprise side, spending has been driven by the need for efficiency, resulting from the age old mantra that 70% of IT spending is on maintenance, with only 30% on innovation (as noted by IDC). Going forward, we see several magatrends, including the acceleration of virualzation, which crives a move toward the private and public clouds, as well as the acceleration of mobile computing (smartphones and tablets). Both these trends, in our view, will cause Unprecedented data growth and in turn increased demand for IT infrastructure. More important, the analysis of such data using software and services will alow corporatons to adjust their business strategies accordingly. In other words, IT spending will become less about efficiency and more about innovation. This will drive @ shift in the sector toward all, things smart: smart devices, smart software (for analytics), and ulimately smarter Investing by enterprises. Such @ shift comes at @ time when we would argue that there has been relative Lunderinvestment in technology, This means that a healthy corporate backdrop should drive growth in the IT sector. As shown in Figure + our top-down and bottom-up analysis, suggests that the IT sector will see 4-5% growth in each of the next two years and 4% longer term. From a device perspective, we believe that the innovation in smartphones land tablets will accelerate the move toward mobile computng and drive healthy value ‘growth of 14% and 65% in these segments longer term. To compete in this increasingly complex consumerlcorporate world, traditional IT vendors not only need to sustain hardware innovation, but equally need to master software and services as a discipline, Figure 1: We Expect the Global IT Hardware and Services Sector to Grow 4% LT Driven by Strength in Tak Services, but Partially Offset by Weakness in Traditional PCs, Servers and Printing USS. dillon, ulossetheraice sated lots and [ Haraware (USS bn) 2007 708-2000 2070 _-B0TE—-20TE ——a1SE CAG TO-T8E ‘Consumer PCs ex able 1 tor iore ist 997 818 aa 32% Wenange 176% «105% ORK 13H 8H Commersal Posextablls M21 4717S 1282 taaa RA a2 65% Seehange 12a 99% ATI 48TH Tablte 00 00 0038 242 1% ‘enange a NM 242.7% 628% GobaiPcs Baa bad 2100 aaa 2593 2714 FT 73% hehange Meh «88 BK TSH AR 48H Servers ssa 827 378A 480 «9 2.0% Seenange 49% 48% 48a 182K 22H Storage 7200 2A 07 BS 303 20% ebange 63% th 8.2% BTH THB Printing ws20 1352228820 8s oa% % change NM 25% 81% ITOH 8H services 74306085 760.1 720815350 eos 49% % change sore ase 1% 25m ae st Total Hardware Tisoe 42500 aaess izaa3 427084788 A080 7% ‘wenange NM 62% 7% «OBO Ssmartonones aut 4848S tag tS wei sa ‘ehange NM 05% 240% 65.2% see 184s Sowce: Garner, Company daa, Creat Suisse estimate. ‘So what does this mean for an investor focused on the IT hardware sector? In Figure 2 Wwe list the top ten key conclusions forthe sector based on our extensive analysis. These highlights combined with valuation lead us to our current recommendations. Our high-convietion, Outperform-rated stocks are Apple, EMC, and HP, while our key Underperform-rated stocks are Dell and Lexmark. We inate with a Neutral rating on IBM NetApp and Xerox.Carprrsuss™ Apple—The Most Valuable Company in the World? \We conclude that Apple's competitive advantage in its software and hardware combined with the momentum around its ecosystem should be able to deliver outsized evenuelearnings growth of 50%/45% in FY10-12, significantly ahead of expectations (around 24% higher for FY 12). Our EPS estimates for FY 1/12 are $25.11832.49. "iPhone stil the driver. Actoss the key smartahone success factors of software, services, product portfolio, dstrbution, brand, IPR, and chipset efficiency, we believe, three years after the launch of the iPhone, few competitors have managed to narrow Apple's advantage. This means within this fast-growth industry (smartohone unit growth of 5254/92% in 2011/2012), Apple's smartphone share should continue to rise to 20% in 2012 driving volumes of 72mn!1 12mn in FY1112 with revenue of $47Hn and $67bn 1 iPadmaddressing @ $1206n market LT. Our proprietary analysis for tablets (takes into account factors such as regression analysis for long-term computing demand, pricing by tier, and cannibaization of multiple industries) highlights thatthe tablet market could re to $1206n by 2015. Within this segment, we believe Apple will dominate, given aggressive pricing, time to market advantage and a software edge, maintaining share as high as 50% long term. This means that iPad should become a $34bn business by FY'12 Further, our proprietary BOM analysis implies that GMs for this business will expand to 35% by end FY‘ from around 27% levels seen in FY10. "= Sit room for an extra $10 in EPS. We believe that a low-end iPhone, greater push into ‘emerging markets, or enterprise traction could add $10 of EPS, Even beyond this, we see scope for Apple to leverage its ecosystem and its current installed base of 200mn (rising to 700mn over coming years) with revenue from advertising, broadcasting or perhaps the TV business, © Valuation. We arrive at our $500 target price using @ combination of PIE, DCF, and HOLT analyses. On our CY‘2 estimate, Apple trades on a P/E multiple (ex-cash) of 89x, which we believe is inexpensive, given the potential for earnings growth of 48% lover the next two years EMC—Virtualization and Cloud Drive Growth ‘Our positive outlook for secular end-markst growth and strong competitive positioning should drive 20% earnings growth per annum. Our pro forma 2011 EPS le $1.50 and our 2012 EPS of $1.83 is 7% above the consensus, and we see 30% upside potential '§ Share gainer ina fast-growing market. We expect the networked storage market to {grow at @ 10% CAGR in 2010-15, drivan by explosive data growth. Second, within this market, we beliove EMC will expand its leading 26% storage market share given i ranks frst on our proprietary storage vendor scorecard (this was also well supported by the recent Credit Suisse IT Survey). We forecast incremental share, driven by gains in the midrange and low-end to drive robust double-digit revenue growth for several years, Long-term guidance of 1996+ revenue growth looks conservative for three reasons. First, we see continued success in backup and recovery (Data Domain and Avamat ‘combined are now over a $tlniyear business). Second, given virtualization as a trend seems to be accelerating, this bodes well for VMware and presents an inherent strategic advantage given EMC's ~80% ownership. Third, we see the potential for solid traction fr the recently closed Isilon acquisition, VPLEX and Greenplum solutions. = Operating margins: scope for modest expansion. We forecast OMs ising to 22.9%124.6% in 2011112 feom 22% in 2010, riven mainly by stable gross margine ‘owing to a software rich solution, richer mix, and SG&A leverage, "© Valuation: 30% upside potential. Our target price of $34 is based on @ combination of PIE, DCF, and HOLT® analysis and suggests upside potential of 30%. The implied P/ECarprrsuss™ multiple of 16x (ex-cash) is warranted, in our view, given strong visibility on growth and impressive free cash flow dynamics (consistent FCF conversion of over 100% over the last 5 years). HP—A $6.00 Print We believe HP's transition from @ hardware-centrc business model continues, with services, storage, and networking rising in the mix and driving group margins higher. This should enable the company easily to deliver $8 of EPS in CY12; therefore, we see Ccompeling upside potential of 45%. (Our EPS estimates for FY 1/12 are $5.36/$5.80.) = Services (99% of O)) resuming growth with scope for margin expansion. Despite the recent hiccup, we see services growth ahead. Our proprietary IT services demang forecast suggests thal when compared fo GDP growth, hardwarelsoftware attach or fo corporate profits, global IT services are being under-consumed globally, providing a backdrop for accelerating revenue growth. Adaitionaly, we see HP gaining share sven postive exposure to faster-growing emerging markets and a robust end to end ‘offering. Finally we believe current cost saving plans will drive FY11/12 segment ‘margins to 15.594/16.0% and to over 17% by FY'14 (15.8% in FY10). = ESSN (20% of 01) benefits from networking. Within ESSN, even allowing for share declines in servers, we believe that end market growth in storage, combined with the ramp of networking should drive sustainable margins. = Three moves the now CEO should consider. (1) Re-invigorate the software segment \which contributes only 5% to profit and is a tenth the size of IBM's software business, {through hiring and M&A. Here the company’s cumulative FCF of $85bn over the next five years offers flexibly. (2) Drive an all-out price war to socelerste share gains in networking at Cisco's expense, since segment margins, though lower than Cisco's, are materially accretive to HP's 24% gross margins. (3) Consider exiting the PC business in light ofits commodity status and limited strategic benefts, = Valuation—compelting upside potential, HP currently trades at 2 PIE of 7x our CY12, EPS versus peers DelVIBM at Ox/11x and relative to the stock's five-year historical multiple (12x). Our $60 target price is based on a blended average of PIE, DCF, ana Credit Suisse HOLT” analyses. IBM—The Big Blueprint IBM is successfully pursuing a combined hardware, sofware, and services strategy that, n four view, serves as a high-quality blueprint for the entire IT sector. This strategy should drive aggregate market share gains and allow for sustainable FCF of $15bn+ par annum, driving EPS of $21.95 by 2015 (above company guidance). However, we believe shares curently reflect this potential growth, and hence we inate on IBM with a Neutral rating and TP of $175, (Our EPS estimates for CY11/12 are $13.21/ $14.88.) Services (40% of Profs)-Well Positioned within @ Recovering Market. Our services fend demand model shows that, when compared to GDP, hardwarelsoftware sales, and corporate profits, IT services are being under consumed by the global economy. This suggests accelerating services revenue growth in coming years. Given @ strong ‘emerging market bias and a robust end-to-end oering, we believe IBM will gradually capture share. For PTI margins that have doubled in recent years to 14%, we see only ‘gradual expansion, given changing deal structures (smaller), and rising competitive dynamics especially rom offshore players, which wil largely offst efficiency drives. AS such, we model 18% margins forthe Services business by FY‘. © Software (48% of Profits)-an M&A Machine that Could Drive Upside. Our analysis, suggests that software acquisitions (i successful as in the past five years) could eddCarprrsuss™ '$3 to long-term EPS (by 2015), making IBM's current projection of $0.90 of incremental EPS from acquisitions conservative. "© Valuation, With over 50% of incremental EPS coming from M&A and buybacks over the next five years, we believe the quality of growth is somewhat low. As such, while at {a PIE of 11x our C¥12 EPS, wo believe shares do deserve to trade at a discount to the SAP. A blended average of PIE, DCF, and HOLT® suggests fair value of $175, implying only 8% upside ‘rom current levels, and hence our Neutral rating. We woula become more constructive below $150 or with increased evidence of robust organic ‘growth Xerox—Turning the Page to Services (While the transformational nature of Xerox's acquisition of ACS in 2010 highlights the long-term secular issues in Xerox's legacy printing business, we believe it was @ smart ‘move, as It positions the company in a (relatively) faster growing market. However, givan limited upside potential of 16% to our target price, we initiate coverage with a Neutral rating. (Our EPS estimates for CY11/12 are $1.10/ $1.30.) = Services (45% of revenue)-the growth segment: watch the synergies. Our Global Services demand model shows when that compared to GDP, hardware/software sales, ‘and corporate profits, IT services are being under consumed by the global economy, ‘This will rive accelerating revenue growth in coming years. Within this context, we see faster growth in segments such as process management (62% of Services segment revenue) and slower growth in hardware maintenance and support (13% of Services ‘segment revenue}. In Services, we note that strong execution on synergy targets could drive upside to estimates. © Technology (48% of revenue)-muted market outlook offsets company positioning. We forecast the printer hardware to be flat and supplies market to grow at a 1% CAGR between 2010 and 2015. While the outlook for the end market is indeed uninepiring, Xerox's exposure (limited to the laser segment), enterprise focus, and ranking on the proprietary Credit Suisse printer vendor scorecard suggests that the company is positioned to maintainlgrow share. Further, @ lead in Managed Print, which is @ secular ‘growth area within printing ($8.5bn market in 2040), is a bright spot 1 Valuationsimited upside. Our price target based of $12, which suggests about 16% upside from current levels, 's based on a combination of P/E, DOF, and HOLT* analyses and implies a PIE multiple of 8.2x(inine withthe 2-year historical discount to the market multple) on our 2012 EPS of $1.30. Increased visibilty on long-term synergy targets and evidence of faster organic growth would make us more ‘constructive on Xerox shares. NetApp—Solid Growth, Limited Margin Leverage ONTAP Despite our positive outlook for end-market growth and the company's competitive positioning, we initiate coverage with @ Neutral rating, given our view that limited operating ‘margin leverage will cap midterm earnings momentum, which we think is erical for shares trading at a 22x NTM P/E multiple. Our pro forma FYIVFY12 EPS estimates of $2.05182.38 are 0%6/4%, respectively, above the consensus. 18 Share gainer in a Growing Market, Relative to our expectation for an 8% LT CAGR for the storage market, we expect the segments that NetApp is focused on (mid-range) to ‘grow at a faster 13% CAGR. This positions NetApp well, given over 70% revenue exposure to this faster-growing segment. This view is reinforced by a second place ranking (afer EMC) on the Credit Suisse storage vendor scorecard, which ranks vendors across eight matrics we believe are important in the storage market. WeCarprrsuss™ ‘expect long-torm share to expand to over 17% from about 10% in 2010, and as such, we model robust FY12/13 revenue growth of 20.5%/22.3% = Operating Margin Lovorage is Capped. While NetApp Is inherently a mid- 208 ‘operating margin business (versus ~19% in FY11), the company is investing to capitalize on current momentum. We believe this is the right strategy: however, this caps margin leverage NT. In FY12, excluding any impact from Engenio, we expect ‘operating margins of 19.3%, on steady gross margins (owing to a software-rch offering) and spending growth, "= Valuation. Our target price of $54 is based on a combination of PIE, DCF, and HOLT® analyses and implies a P/E muliple of 20x our CY12 EPS of $2.70. This reflects a healthy premium to the market (toward the high end of the fle-year historic range). While @ promium multple is warranted given growth prospects, limited scope for midterm earnings momentum makes it challenging to argue for an even richer premium, Dell—A Long Transition to the Enterprise We believe that Dell is in the midst of an ambitious transformation to become a more strategic enterprise player; however, we are concerned that this will take some time. Meanwhile, we see limited structural organic revenue growth, wth only gradual scope for futher margin expansion, and consequently, a lack of bottom-ine growth. (Our FY12M13 EPS estimates are $1.65/81.63.) "© Three issues against long-term revenue growth. Fist, we see the PC industry, excluding tablets, decining 6% long term. Second, within this segment, we believe that Dell may ‘continue to gradually cede market share, owing to a weak emerging market position and distrbution, Last, owing to Vitualization and increased competion potentially, we see: Del's sorver business soeing flat salas at bast. Even allowing for grow in storage and ‘services, the long-term biended revenue growth ie lat (down 0.2%), = Gross Margin ~21% long term at best, Based on out analysis of grass margins by ‘segment, we believe there will be expansion from 19.1% to 20.6% long term, driven by improved mix. Given limted levels of further opex improvements, we estimate OMs of 6.8%616.7% in fiscal 2012/13 = In need of a transformation, but lacking the financial frepower fo compete. Simply put, Dell needs a transformation, and the issue is that this wil involve M&A in the software, Services, and storage areas. Here, not only is the curent track record mixed, but the ‘company will also nead to compete for targots with peers such as IBM, Hewlett-Packard, ‘Oracle, and Cisco that have deeper packets, and more attractive platforms, = Valuation, more of a relative Underperform. Our price target is $18, giving limited downside risk based on OCF, HOLT", and PIE analysis and making Dell a relative Underperform. If the stock reverts to is historical multiple of 8x, the downsige is $15 implying limited absolute downside. Lexmark—Focus to Return to Secular Challenges (Our 2011-12 EPS estimates of $4.44 and $4.18 are 6% and 12%, respectively, below the consensus. We expect negative PS momentum through the year to tive Underperformance, as cyclical talwinds that the company enjoyed in 2010 fade and as secular issues related to the challenged outlook for market growth and the company’s positioning relative to peers come back into focus. Consequently, we are inating coverage with an Underperform rating and a $35 price targetcarorrsuss™ "= Focus to shift to secular market challenges. We expect the printing hardware and ‘supplios market to grow at a 0%/1% CAGR through 2015. On hardware, our outlook is driven by a fading refresh cycle near term, while on supalies (the main driver of profits), ‘our view is driven by a shrinking installed base. We expect all the growth to come from developing regions. © Declining hardware and supplies revenue. Given limited developing market exposure ‘and a fourth-place ranking (of four vendors) on the Credit Suisse printing vendor Scorecard, we expect hardware revenue to decline. in addition, despite higher usage, driven by a focus on laser and business inkjets, we expect @ shrinking inst drive dectining supplies revenue. led base to "© 2012 eamings power closer to $4, nt $5. Declining revenue growth will rive operating margins lower, in our view, with risks to margins being to the downside, as more investments are likely as the company addresses end-market concerns. As such, we ‘expect the 2012 consensus estimate to move closer to our $4.16 estimate as the year progresses. "© Valuation ~a relative underperform. Our target price of $36 is based on a combination of PIE, DCF, and HOLT? analyses and implies a P/E multiple of 84x our 2012 EPS of $4.18. We note that over the last 2-years, Lexmark shares have traded at a 25%-35% discount tothe market, rardware 8carorrsuss™ tomva sor Ten Themes for the Next Five Years In this section, we lay out our long-term outlook forthe $1.2tn IT industry (which includes ‘communications equipment, mobile computing products such as tablets, etc). Clearly, in such a significant industry, which accounts for ~2% of global GDP, the dynamics of supply, demand, and competion are especially complex in a backdrop of convergence and consolidation, as evidenced by recent M&A activity. We have attempted to develop a framework to project demand and think about which vendors are best positioned to caplure value within the industry over time. As such, we arrive at ten important conclusions: Figure 2: Ten Themes for the IT Industry 1] Gycial tlds Yr now, Our analyea oT naToeconom and corporis enwTonneat combed wit our propraanyT Suny structural #6 growth suggests that toll T meusty (PCs, server, storage, pong and een/ieos) will enanue fo oo cyical Fecovery in 20112012, with dusty rovenves growing to 81 tr, mplyingarouné 45% growth 2) Compute demand srves Our aderearabv market analyls or ables (based on PC orecans By price pan) suagosta ha he bat tablet market to $1200n long met could veech 298mm uns by 2015 (42% af lll PC uns), wi Favenues of $1200 3) Smatoranes sat cross Based ov our pos ear aTOTSaUI Svan we oaimnae Hale sparohove Marketa grow Far ‘on unitmarkby 2015, 287mm uns in 2040 lo Sa4mnnt.O4bn unin 2042!207S,mpling a GAGR of 20 5% over te Nex Fue years 4) Siorgo has several secular Orven By accslaraing unsvudlured dala growth, server vtualaaion, and regualoneomplance rome drvers| Fequcemonts we bohove storage will become ionsing'y iporiant. As ech we expect sora Uarawar sofware, and sarees reveruo to grow ata healthy 7% por anrum to Sas By 205 1) Services eet fora racval Based on oor view hat sriena ie being underoonsumed lobaly, we are ooking fr around 8% top-ine recovery ‘row forthe sorices mare (87800r= In rowan n 2010) over ne not fe years 3) Lackusir au servers, We aee muled oulaok for savers and WadiionalPC martes, aa wa oxpee revenue docings al GNGR@ af trectoral PCs, printing 2% and 8%, rospecvaty. Fo the pine market. wo oxpec reveruo fo sfow only 1% In 2010-2015 owing to Shaking insted printer bes 7) Race changes to ha PE tonal PC vendor, value enac? ‘ot vendor ike Meroaof and intl ‘wa come under pressure inthe PC make! 8) Competition for With an accelerating end lo become an endo end solion provide, we ae eeing signs of tadional datacenter teatngup part ting head to Nea, wih Cisco's push no the sever market, Oracles move inthe Eorver and storage mark's, 7d H's ofr inthe networting spac {B) WANT cerise for Our analy fr gros cash ot cae, and RED svestronts atop 30 ecMTaTay Compan Over Te onindod consolation Suggouts that M&A acti ae Ike lo conteue in the sector, wh IBM. MP, sid Del kay to bo most acive 10) Tho coud may otbe as Recor announced cloud oferngs For 18M, HP, and Dsl are fundamental snl f infastrucure Incremental you tink ousouringvervices alway being provided by these companies. As such we expet Ie inpal of cloud may note s incremental as porceved y he market “Source: Company dala, Cred Suisse estimates, Nola Il cover by Credit Suisse Samiconducorarajal lab Pilzr, Mirosot & Oracle covered by Cred Suisse Sofware analyst Pi Winslow, Cisco covered by Cre Sulsse Communicaton nrastructure analyst Pau! Sivertein 1) Cyclical Tailwinds for Now, Structurally 4% Growth ‘Much of our analysis deals with the microeconomics of the IT hardware sector in terms of the fundamental demand rivers of servers, storage, services, and PCs, However, for an Industry that represents nearly 50% of U.S. investment in private fxed assets, the ‘macroeconomic environment is a key consideration, After economic factors such as GDP. growth, corporate health, CEO confidence, and other leading indicators are taken into account, we conclude that there isa robust economic backdrop that lays the foundation for {growth in IT spending atleast trough 2011. This wil likely be a contnuation ofthe cyclical recovery seen in 2010, and we base our conclusion on several factors rardware 9Carprrsuss™ “16 March 2011 Figure 8: 66 Investment of GDP Figure 4: G4 Nenfinanclal FCF % of GDP ‘A conducive macroeconomic and corporate backdrop. As depicted in Figure 3, the last few years have seen an unsustainable reduction in fixed investment as a percentage of GDP; roturning to trend alone will result in a positive uptrend in tech spend. A combination of strong balance sheets, cash at record levels (in the U.S. at 6% of assets), and robust FOF (at 4% of GDP, the highest levels since 1995), combined with increasing levels of business confidence, means that corporations have the continued firepower to invest for growth; this is noted in Figure 4, Figure 5, and Figure 6, Further, Gé business investment stands at 16% of GDP (Figure 3), one of the lowest levels on recore, yet tech accounts for a significant 40%+ of investment spending. The macroeconomic drivers behind corporate discretionary spending are very appealing and will prove to be supportive fer IT spending inthe coming years. Figure 5: U.S. Nonfinanclal Corporates—Cash as a% of Figure 6: U.S, Nonfinaneial Corporates—Cash as a % of Total Assets Market Capitalization Tech is being underconsumed. Having analyzed levels of technology, net investment relatve to GDP, long-term trend, and depreciation, we find that levels of tech investment remain structurally below long-terms levels, even alter the recovery in 2070. (See Figure 8), For example, net tech investment (capex minus depreciation to GDP) normally runs at an average of 0.62% with a peak to trough of 1.4%; however, it currently languishes at only 0.1%. This relative underconsumption of tech in major economies adds to the argument that spending should continue to recover in 2041. In fact, net tech investment i 2009 for IT haraware (Figure 7) was negative, implying that assets were depleted. Given @ Similar recovery in 2002, and combined with our proprietary IT forecast models, we aro predicting 8% IT capex growth in 2071 and 6% in 2012, bringing us to prerecession 2007 levels of net technology investment as a percentage of GDP (dashed line in Figure 8). thermore, as we discuss in our macroeconomic analysis, otner similar forwara Indicators, such as CEO confidence, corporate profs, the ISM index, and durable goods, corders imply positive momentum ahoad. rardware 0Carprrsuss™ Figure 7: Investment in Hardware Turned Negative In09 Figure 8: A Strong Potential Bounce In Investment rel investmantin IT hardware and stirs a eof GOP rotinvestmentin rat 9 91 GOP “Source: BEA, Groat Suisse extinios “Source: BEA, Cod Suisse estimates. Our proprietary Credit Suisse IT Survey is supportive of continued cyctcal taiwinds. Based upon our survey of top IT dacision makers at major global corporations, we find that some 63% of respondents expect IT spending to increase, with a consistent response across all regions. In aggregate, IT headcount growth is expected to grow some 6%, with spending increasing at significantly faster rates than in 2010; the fastest growth will be seen in areas such as services, storage, mobile devices, and software. Figure 8: Credit Suisse IT Survey—What Are Your Year-Over-Year Growth Expectations for Your IT Spend in the Following Areas in 20117 Storage Senices Mobile Devees and Servees Pc Security Networking Soners 2.0% 1.0% = 01.0% 20% 8.0% 4.0% ‘Soure: Great Suse 17 Survey, Pabrary 207% Boyond tho cycle, a 4% growth industry. Beyond the cyclical component ofthe recovery in 2011 and 2012, we believe that revenue growth will be around 4% per annum, rising 0 8 $1.3tn industry by yearend 2012 for the IT sector (PCs, servers, storage, printing, services). We note that each of our industry models are built upon specific proprietary analysis rather than simply looking at historcal levals of revenue growth. Within the segments, we believe that attractve growin segments are tablets, smartphones, and IT storage; we see muted revenue growth for traditional PCs (ex tablets) and servers,Carprrsuss™ 2) Compute Demand Drives Tablets to $120bn Market \We fundamentally beliove that PC projections that are based on forecasting seasonality for desktop, notebook, and netbook demand are flawed and subject to excessive volaity We have developed what we believe is @ unique allemative to forecasting demand for consumer and corporate compute power (all product categories): Consumer PC growth of 17% (including tablets) long term. We have developed what we believe is the frst econemetric model for consumer PC demand using cross-sectional analysis across 42 countries and based upon over 1,000 data points. We find that there is, 2 statistically signifcant relationship between PC affordability and the PC penetration per capita, (with R-squared ranging between 71% and 86% from our multiple regressions). Based on these fundamental relationships, we demonstrate that the elastic of demand remains above 7.0. Simply put, this means that a move to lower price points will drve incremental volume. Also, based on an extrapolation of product teardowns, we demonstrate that an average quality low-end PC is plausible at a $200 ASP within the next five years. in turn, what this means is thatthe installed base for PCs wil ise to 1.2bn from ‘680mm last year, with consumer PC volumes growing at a 17% CAGR (2% exctablets) LT. Commercial PC volumes ta show robust growth in 2011-12, driven by @ corporate refresh, \We estimate that the average age of the installed base currently is six years, which to us suggests that replacement volume wil recover in the near term. Furthermore, our proprietary Credit Suisse IT Survey (polling 60 top IT decision makers at global fms) Suggests a further boost, given the transition to Windows 7 and new hardware releases {more powerful specs and chip releases). As PC penetration of the labor force continues fo Increase, we forecast commercial PC volumes to grow 17%/16% in 2011/12 (139%)10% ‘eetablets) and 11% long torm (5% x-tablets}. Tablets are aifferent for many supply and demand considerations. The tablet market Is inherently challenging to forecast, given its recent introduction. From a fundamental point of view, we do believe tablets are different, and there are several reasons why tablets wil have a mote meaningful impact on the PC industry than netbooks historically. These Include: (1) optimization for megia consumption, (2) numerous applications (app ecosystems), (3) hundreds of indusiry-specific uses (restaurants, healincare, education 2%), (4) mobile operating systems that are optimized for the smaller form factor, and (5) leverage from new distribution channels (carters) and (6) instant onflonger standby. Tablets represent a $120bn market long term. We acknowledge that tablets are unlikely to replace all computing needs. Far this reason, we adopt a price point-based approach that, assumes that demand for a given level of computing necessity can be approximated by price level. We use our global PC forecast by price point to determine the addressable ‘market for tablets; we conduct a penetration analysis st each of these tiers. For instance, & low-end PC for the consumer market at $300-499 can be better served by tablets (63% LT) versus the high-end $1,000* category (wo assume 0%). Likewise, we perform a similar application-based approach for the corporate market. We conclude that the tablet market could represent @ $120bn market by 2015, with units reaching 298mn (or 42% of total PCs). Furthermore, while we believe that over 50% of tablet volume will be consumer based, we are surprised by the degree to which corporate adoption appears to be taking hold; this is further conficmed by results from the Credit Suisse IT Survey as seen inCarprrsuss™ Figure 10: Tablet Adoption—High Enterprise interest Figure 11: 30% of Commercial PC Demand in 3 Years? ‘Question: your company curently deploying tablet devices, or might Question: What % of your global employee base havelwillhave @ you consider depioyng lable devies sometime inthe next 5 years? lable deve (supported by enterprise) atthe flowing pont in ine? 3) Smartphones Set to Cross 1bn Unit Mark by 2015 Based on our proprietary model, which takes into account the total cost of ownership (TCO) for a smartphone, income distribution, and penetration of the addressable market wwe conclude that the addressable market for smariphones could be as high as 2bn longer term. We define TCO as the upfront cost that a consumer pays for a smartphone ‘combined with the annual service cost for a basic voice and data plan associated with that device, Our smartphone model suggests that by 2015, the global smartphone subscriber base will reach 1.9bn, (ie., 98% of the 2.0bn addressable markel). Based on this long-torm estimate, we believe that smartphone volumes will grow from 297mn in 2010 to '594mn/1.04bn in 201212015, implying a CAGR of 28,5% over the next five years, Despite seeming optimistic, our cannibalizaion analysis and handset price point work sill suggest thatthe risks are to the upside. Figure 12: Smartphone MarkstOur Long-Term Forecasts Are Based on Thre Different Methodologies ‘Analysis Mothodeloay Result Implication Total Goat! Propratary Gradt Guase model hat 2 ibvadarssable waratlar By 2018 we eatmate 2.000 people tbe Ownership” forecasts the smartphone market using the smarphanes by 2015, up fom patente smartphone subseber, ana we CO (total cost of awnerenp of rene 2009 fesume 1 sbn amartphone subeerber, which ponetaton of the aderessale market market. Hence, we Seleve ur emarghone Subscriber estimates could prove conservative ‘Gapnatzaton of Farecasing sharphone maiialsaaed on G38 smarphone us Oui} Odbn esimale or smaripnone waumes By consumer incremental opportunity to cannibal her 2015 based on smartphones 2016 iooksachivable based on out Suen 2s MP3 cannbalzing ater CE device cannalzaton wor: this suggests thet fevioes Blayers, geming consoles, PNDs, carers, segments. fmatghone volumes cols be a8 high as 9880" Sand midena phones {OW resucon — Looking w emariphane nariet based on 1 Ofer smarphone ania h The 1Obn srarhane ont rune by 07518 fand pe band blat-materasraducton fore midend” 2015 based onthe 8OM again nlne wih our publanad volume forecast anayses smarghone and how auckysmaiphones reduction ana price band based on TCO an afordaity ana ‘ou pantrate lower priest anateeCarprrSunss’ 4) Storage Has Several Secular Growth Drivers \We project the storage hardware, software, and services market ($47bn in 2010) to grow ‘at a healthy rate of 7% per annum to over $85bn in 2016. In particular, demand for storage capacity is being driven by accelerating unstructured data growth, driven by the explosion of digital content, server vidualization, and regulatory and compliance requirements. As such, we expect storage will become increasingly important, not only as enterprises cope with rapid data growin, but lso as they continue to optimize their vitalization Implementations. Figure 4 itual Machines: Increasingly Contribute to Unstructured Data Growth 108 om on 0% a —f aK 108 ow yee Spe syer [EiGc ecredan IDC union dn GCS va nasine ‘Sour: IDG, Croat Suisse IT Survey Noe: IDC year CAGR extrapolated rm 4 years Within the storage hardware markol, Here continues to be an ongoing secular shit rom direct attached (to the server) architectures toward networked storage architectures. This means that networked storage (SAN and NAS), which accounted for 83% of storage hardware spending in 2010E (up fom 62% in 2005), wil rise to 92% of storage hardware spending by 2015E. Even beyond growth, we believe the market represents an attractive ‘and strategically important segment, as software and services collectively account for 58% (of the overall storage market opportunty. In turn, ths explains the frenzy of MBA activity in the space by all major vendors, inckuding HP, Dell, IBM, NetApp, and EMC over the past five years, 5) Services Set for a Gradual Recovery While companies ike IBM, HP, and Dell have historically been viewed as hardware Companies, the actual end markst for IT services dwarfs the hardware market opportunity, {8 services is over a $780bn market, some 2x the value of hardware procured. We Conclude that the macro backdrop for IT services is positve inthe long term, Over the next five years, whether we look at IT spending relative to GDP levels, attach rates to hardware ‘and softwere, or corporate revenue forecasts, (based on S&P forecasts) we find that IT services are being underconsumed globally; this points to accelerating growth ahead. For ‘example, global IT services as a percentage of global GDP are 1.3%; however, in 2000- (09, it has risen consistently from 1.07% to 1.32% and hit a high water mark of 1.37% in 2008. If this trond rasumes, then the CAGR in the services market would be 5.8% por annum, (We are forecasting slighty lower annual growth of 4.9% during the same time period.) Within this context, we see faster growth in segments such as process management and slower growth in segments such as hardware maintenance and suppor. “4Carprrsuss™ Overall, after imited revenue growth in IT services in 2010, we believe that 2011 wil Continue to show late cycle characteristics. 6) Lackluster Outlook—Servers, Traditional PCs, Printing Servers Anemlc Server Market Revenue Growth Ahead We believe that overall server market revenue will decline at a 2% CAGR in 2010-15, declining to $44bn by 2015. In particular, the x86 market (98% of units and 65% of revenue) has enjoyed strong growth in 2010 (sales were up some 23% in 2010). Ths is, attributed to both a cyclical recovery and a very substantial x86 processor redesign. Going forward, we expect revenue CAGR of -2% in 2010-15E, with the 8% long-term growth in the 2way under $2,000 category (owing intemet traffic, Web-based applications, and high-performance computing workloads) being more than offset by a 3% long-term revenue decline in the 2-way above $2,000/4-way and above categories owing to the Impact of server virualzation. We would note that, in the latter category, lower volumes ‘are mitigated by higher overall ASPs, which mitigate the revenue decline for Vitualzation-focused systems, PC Ex-Tablets—Lackluster Prospects Given our view that tablets wil represent 42% of computing needs longer term, unit and revenue growin in this segment will be substantial. However, excluding tablets, we believe the traditional PC market will only grow a mere 3% in unit terms and will actually dectine 6% in revenue terms to $172bn in 2048 from $238bn in 2010. We expect particular weakness in desktop and netoook shipments, Simply put, we believe that tablets will be able to addrass lower-end computing needs, especially in the consumer market Printing—a Flat Market \We expect that the $511on printer hardware market (last year) will ee flat growth long term. lowing two years of dectining shipment growth (a 6% decline in 2008 and a 14% decline in 2009), the printing hardware market benefited from a refresh cycle in 2010 that resulted in overall shipment growth of 8% and overall revenue growth of 5%. Within this, we believe the shift to multifunction devices, and away from single-function printers. ang standalone copiers, will continue. Overall, wa expect flat MFP and single-unction printer revenue growth long term, while revenue from single-function devices will remain fat ‘through 2015. Our outlook for the $73bn supplies market that generates the bulk of the printing industry's profits is equally muted, We expect supplies industry revenue to also grow at a 1% CAGR from 2010 to 2018 owing to a shrinking installed base, which will, offset more pages:printed-per-device. 7) Radical Changes to the PC Value Chain? [A clear conclusion forthe entire PG value chain is thatthe combination of strong tablet unit growth and revenue growth wil come largely at the expense of traditional PCs. Combined with the share gains by Apple and the potential success of Android, this means that the lraditional value chain in the PC indust'y may see radical change ahead, An important supply side consideration is the deficiency that currently exists in the Intel and Microsoft platforms for tablet use. Owing to innovation from platforms including Android, the tablet ‘market has litle dependence on the Wintel platform, and will have a more competitive processor market. Indeed as shown in Figure 14, this platform accounts for 50-80% of value and has done so consistently over the past several years,Carprrsuss™ “16 March 2011 Figure 14: Wintel Operating Proft Share ofthe Industry Has Been 66% on Average ‘Scwee: Company data, Creat Sues estimates. ow value share will evolve in the industry is hard to predict; however, we believe that vendors such as Apple or HP are attempting @ more vertically integrated approach to tablets, Traditional smartphone vendors including Samsung and Motorola are aligning themselves with the Android platform. The tablet markat's share of the computing value chain may eventually become similar to the smartphone market. The contrast ie staring whether the actual branded hardware vendors vertically integrate or otherwise, they may have some 90% share of value inthis industry. In the PC market, the equivalent portion is 20%. rardware 6Carprrsuss™ Figure 15: The Operating Profit Share of the Smartphon Nonexistence of a Chip/Software Ouopoly (Wintel Industry Is different than the PC Industry, Given the ws 1th tr ony _ om F aise saan nen & aust Ae sven, Merosatalon Bon sy i _ Savoy eso tone o 30% 10% of ndustry profits, vs. over 40% for ‘Scare! Company data, Grea Sues estate. \We would highlight that, from the Wintel perspective, the previously mentioned dynamics are only confined to the tablet portion of the market. Clearly, there will be a significant portion of revenues and profits for which the platform will essentially remain untouched. ‘The alternative argument ‘To be clear, CS Intel analyst John Pitzer and Microsoft analyst Phil Winslow have a more optimistic view for the competiiveness of both companies in the PC market, but specifically for their potential in the tablet market based upon several factors Improvements coming from Oaktrail and Medfeld. intel has demonstrated several Oak Trail (82nm integrated SoC for Windows) based 8-10" tablets and gaming consoles to launch in TH11 from several OEMs (Dell, Samsung, Toshiba, Acer, Asus) with multiple OS (Windows 7, Android, MesGo). The company have recently reported 35 tablet wins with the fist Androis based tablets in 3011, John Pitzer believes that Intel will improve power ‘consumption characteristics of jis products, focused on the tablet market at 32nm (products expected in 2H) and further in 2012 with 220m ramp for SOCs (Medfield). Windows 8 in 2012. In mid-February, Android Central posted supposedly “leaked” screenshots of Dell presentation slide deck depicting tablet roadmap that targets @ January 2012 release of a Windows & tablet. To date, Microsoft nas not specified & ‘lease date or a detailed development timeline for Windows 8 but noted the traditional 24- 36 month development cycle al CES 2011 thie past January. Given that Microsoft reportedly compiled Windows & Milestone 2 in December, the Credit Suisse Software Team bolioves that a beta of Windows & may be released as early as the SeptemberCarprrsuss™ quarter, and they continue to expect that Windows 8 will be released in 2012. Though the Credit Suisse Software Team expects enterprise adoption of Windows 8 to be considerably lower than Windows 7 given the strong adoption of Windows 7 and the shart release period between the two operating systems, the team views Windows 8 as 8 ‘meaninglul release for Microsoft in terms of the company’s positioning in the consumer ‘market—particulary in the tablet segment. The Credit Suisse Software Team believes that Windows 8 will not only support ARM in addition to x86 but wil also be available with three User interfaces for OEMs to chose from, one of which will be tablet-optimized based on Microsofts Metro Ul, whieh serves as the interface for Windows Phone 7 and Xbox. The Credit Suisse Software Team also expecis meaningful improvements in power consumption and boot time with the release of Windows 8, which are important characteristics inthe tablet market, Therefore, the Credit Suisse Software Team ultimately believes that Microsoft wll have a much larger positioning in tre tablet market than Wall ‘Street currently anticipates, especially in the stil nascent corporate tablet market, Integrated Versus Fragmented Debate, Apple is unique in thet integrated approach - they sell a system in which they own all of the software IP and ertical components of the hardware IP. The opposite business model isin the PC market if Dell, Microsoft and Inte were all the same company. In stark contrast, Apple's competitors in both the smartphone {and tablet markets are forced to choose from multiple vendors increasing the complexity with regards to software and hardware integration. While Google's Android (covered by Credit Suisse Entertainment, Intemet, and Cable DBS analyst Spencer Wang) is becoming the de facto OS in these new markets, he integration of software to silicon is significantly complicated by a highly fragmented market in the apps processor arena. In four opinion the integration issues only become more complex as you move trom smartphone to tablet - from a 4" to a 10" screen, OS and silicon solutions need to be more robust and more highly integrated. In our opinion itis parlly the reason for the lack of 10" tablets coming to market in 4010/1011 in favour of loss robust, and in our opinion, less Interesting 7” tablets. This move towards standards could make Intel more viable in the tablet market. 8) Competition for the Datacenter... Heating Up [An accelerating trend in the sector is the desir for large IT vendors, including HP, IBM Cisco, Oracle, and Dell, to become end-to-end solution providers or sole suppliers for the ccomplate solution within the datacenter, including networking, software, storage, servers, and services. The driver of this strategy Is the vendors’ desire to capture incremental Customer wallet share as data centers are rearchitected for private or public cloud Implementations. As a result, we see traditional partners now competing head to head, most clearly demonstrated by Cisco's move into the x86 blade server market with its UCS offering, Oracle's push into servers and storage with Exadata and Exalogic, and HP's move into networking, While no one vendor has yet mastered the complete stack of hardware, software, and services (as shown in Figure 16), with many of the products and strategies stil evolving, we would highlignt the following in order to determine the impact, fon the technology industry:se morn 2014 creprrsusst™ avian Le i 0 sory costo ona _sejeuse ss ppe19 ep hueduog 200s 12s Suomen sors sang ees uonenyerny uotsks ees Bueno size sse9eeg 80513 suoneandey ess weueseuey erg fuoboe9 a wa so¥wo yey ABojouyoo, OWEN JONPOId /M AuedWOD Jed aJeYS JOEY) ‘ouaBronuog a1oy| BuIAHA S| J2qUodeIeG O19 40} UORHadWiOD :9} 94ND) 19Carprrsuss™ Bundling Versus Best-of-breed, Purchasing Patterns Will Not Shift Over Night Few would doubt the strength of Cisco or Oracle's salesforce or distribution. The question 's whether bundled or best-of-breed solutions are preferred. While the debate around this, is not @ new issue, we would highlight, as noted in Figuro 17 and Figuro 18, that at least from a purchasing perspective, there appears to be a strong bias toward best-of-breed products. In the Credit Suisse IT Survey, some 67% of IT decision makers (Figure 17) noted a preference for best-o-breed versus purchasing from one vendor. Figure 47: 67% of Respondents Prefer Best-of-breed Figure 18: 43% of Respondents Don't Want 1 Supplier ‘Question: you were fo bud for curently have) an intemal prvete Question: In 2074, Ifyou could choose fo purchase the maj of ous, you woul: {Dir hardwarofatwaralerie fom one ono who Would be? aE ‘Sour: Great Suse 17 Survey, February 207% ‘Source: Creat Suisse IT Suvvy, February 2077 Exadata and Exalogic Are Likely to Have Some Impact, [As noted by Credit Suisse Software Analyst Phil Winslow (noted in the reports Dr. Exolove, Part I: Or How I Learned to Stop Worrying (about Sun) and Love Exadata, dated 12 October 2019 and Dr. Exalove, Part II Or How I Learned to Stop Worrying (about Sun) land Love Exalogic Too, dated November 23, 2010), due to Oracle's strength in both database and middleware software and, a growing market share in enterprise applications, the Credit Suisse Software Team believes that Oracle continues to build robust and ‘growing pipeline for the Oracle Exadata Database Machine and that Oracle's appliance strategy—from Exadata to Exalogic—positions Oracle to be a disruptive force in the server, networking, and storage hardware markets. As a result of Oracle's stength in the database layer, the Credit Suisse Sofware Team believes that Oracle is uniquely positioned to increase server performance and lower storage hardware costs through innovation in the software stack ~ this will put pressure on competing hardware vendors, Oracle's Exadata product is offen thought of as the Oracle Sun Database Machine, as it combines Sun database and storage servers, storage disks, and Oracle database software into an integrated system that is optimized for running Oracle's market-leading database sofware, Theoretically, these advantages make Exadata wall suted for OLTP. applications and data warehousing, especially where Oracle database software is the preferred solution, Our survey highlighted several key conclusions on Exadata ‘Significant awareness, but considered expensive. There is significant awareness of the Exadata solution, as demonstrated by nearly 40% of survey respondonts having alroady evaluated the product (Figure 19), but the majorty of respondents on average do not Intend to purchase the product in the near term (Figure 20). Oracle clearly targets Exadata fat the company’s database software installed base, Therefore, because the results dotailed in Figure 20 include both customers and non-customers of the Oracle Database, Figure 20 likely understates the potential demand within the Oracle Database installed 20Carprrsuss™ base for Exadata. Given that convincing a customer of IBM's DB2 database to switch to Exadata represents a much more challenging sales process (given that the customer ‘much migrate its data from one ROBMs to another) as compared with existing Oracle Database customers, who can more easily migrate the sofware to the Exadata platform, While the appliance is on the high end in tarme of price, it is important to nole that does. include a bunale of software, storage, and servers. Figure 19: Exadata—High Level of Awareness Figure 20: .. . Planned Purchases Will Be Gradual ‘Question: Hove you evaluated Oracle's pplance? ‘Question: How kel are you to purchase Oracle's Exadstaepplance nthe tur (unl, 30> very ely? ‘Seure: Cred Suse 17 Survey, Pebrary 207%, ‘Source: Great Suisse IT Suwvy, February 2077 A strong installed base means that it will ikely have some impact on the industry. Oracle will target iis vast software customer base with thie solution, and itis showing early success. At the end of 2010, the company noted a $2bn sales pipeline for Exadata, up from $1bn in mid-2010. The Exadata solution is optimized for Oracle's database software, land this defintely gives the company an advantage in bundling for applications that are heavily dependent on high-performance instances of Oracle's database technology. This Js also true for Oracle's more recently released Exalogic product. In comparison to Exadata, the Oracle Exalogic Elastic Cloud is an appliance that combines 64-bt x86 processors, an infinBland-based I/O fabric, and solid-state storage with the market-leading Oracle WebLogic Server, other enterprise Java Oracle middleware products, and a choice of Oracle Solaris or Oracle Linux operating system software. Exalogic 's optimized for Oracle's middleware and application products, and hence gives Oracle the opportunity to move these workloads to Oracle hardware, Cisco's UCS—Impressive Technology, but a Long Way from Main Stream Adoption. ‘Awareness is high. With its Unified Computing System (UCS) offering, Cisco entered the server market with a broad, flexible data canter strategy that bundles servers, networking ‘and management software into a modular, cohesive architecture that can be managed as 8 single entity. Owing to increased competion from IBM and HP for the data center, Cisco likely fl that it was necessary to enter the server market, as it was losing leverage by not directly controling the server IP. As depicted in Figure 21, in a relatively short period of ime, there is a strong awareness of Cisco's UCS offering, with almost 50% of survey respondents having evaluated i at some pointCarprrsuss™ “16 March 2011 ‘Source: Great Suisse IT Survey, Fobrary 207% UGS, seen as an expensive soliton. A customer evaluating Cisco's UCS products is {generally not only looking only to buy servers, but instead is looking for a comprehensive data conter soluion that includes servers, networking, management tools and storage. Cisco's UCS offering is typically more expensive on an equivalent basis with other ‘competitive offerings, as respondents to our survey noted in Figure 22, Server Solution a as a as 4 as ‘Seure: Gre Suse 17 Survey, February 207% rardware 2Carprrsuss™ Early run rates are impressive. Cisco is experiencing early success with UCS, claiming ‘ver 4,000 customers and an annualized run rate of $650 Millon (nearly Tx year-over-year growth) at the end of February 2011. However, with less than 1% share of the server market, Cisco has @ considerable way to go in order to catch HP, IBM, and Dell in the server market. Cisco has the luxury of building high-performance data cenler architecture from the ground up (with litle hindrance from legacy technology), and although UCS may enjoy a performance advantage, it's widely expected that HP and IBM will catch up in the near term. Although Cisco's go-to-market plan wil likely focus on IT shops that are alraady ‘existing networking customers, those very same customers may hesitate to give Cisco a larger share of wallet owing tothe fear of vendor lockin. 9) M&A—Sector Is Ripe for Continued Consolidation We estimate that the IT hardware, software, and services vendors spent a staggering $5000" on M&A over the past five years. Going forward, we believe, if anything, this trend could accelerate, Figure 28: Gross Cash to 8: y—Healthy Levels of Cash Figure 24: Net Cash to Sales Are Close to Highs Technology companies have the firepower to spend. A glance at Figure 23 shows that the level of gross cash to sales is al 23% (similar to the average over the past ten years). fowever, net cash levels of 8% are higher than the historical average of 6%, as shown in Figure 24. In adaition, we note that FCF levels for the top 50 technology companies are also at record levels, both in absolute terms and as a percentage of sales. Clearly, such signiflcant cash flow ean also be used far a combination of dividends! buybacks; however, given the supportive macroeconomic backdrop, we argue that M&A willbe a priority, 2ssuaren 2081 emery Fer man ‘Seure: Company dala, Credit Suisse research Underinvestment in tech R&D? A glance at Figure 25 reveals that in aggregate R&D intensity in the technology industry continues to decline for the top 50 companies. This was only 6.5% of gales in 2010, which is not only lower than history but also lower than several other sectors. We would argue this lower R&D as percentage of sales points to less internal investment and more potential extemal investments through acquisition, Figure 26: R&D to Sales Ratio for Technology Companies Is at an All-Time Low [R&D to sales rat for help 50 lechrology comanes over time Oss tsa ‘Seues! Company data, Creal Suisse research MBA in technology is on the up. A glance at Figure 27 and Figure 28 shows that some $113bn of M&A was completed in 2010, with 190 deals being completed in the global ‘communications, hardware, software, and services industry. This represented a significant uptick over the prior two years (average of $85bn), but sill below the record levels seen in rardware 24Carprrsuss™ “16 March 2011 2007 ($172bn), which shows an increasing appetite for deals. Interestingly, we also notice that around 70% of transactions (both in terms of deal count and deal value) have been in the software and services area in 2010, Figure 27: M&A Appetite Has S van inerease In 2010... Figure 28. With Software/Servces Being a Focus Area il ‘Source: Company data, Groat Subse research, “Source: Comany dota, Gre Sus research SEES ETT Convergence and the rise of the megacaps. As the technology sector has continued to Consolidate, megacaps vendors, including IBM, Cisco, Oracle, and HP, have continued to got larger. For these companies, there appears to be a move toward a more vertically Integrated industry structure, with bluring lines between hardware and software, and nike the silo approach of the previous decade. Time will tell whether such strategies prove successful. However, what is clear is that, given levels of existing cash, FOF generation, and the intent to drive a bundled solution for the datacenter, consoldation in ne industry is likely to continue 18M, HP, and Dell ae likely to be very active in M&A. Within our sector, we beliove that the mast active companies when it comes to M&A aro likely to be IBM, HP, and Dell Indeed, all three have significant potential excess cash for acquisitions. IBM openly highlights that the company may spend as much as $20bn on M&A in the next five years, ‘mainly in software. We also belive that the incoming CEO at HP may seek to reinvigorate its stagnant portfalio, and Dell is likely to use acquisitions to get to ts targeted enterprise business mix. Even for Apple, we believe that with $60bn of net cash along with a business model and ecosystem that continue to evolve rapidly, if the company is to monetize ts installed base of over 200mn users, M&A may be necessary Figure 29: IBM, Dell, and HP Accounted for 5% of Deal Value in Global Sector Since 2006 USS n llons.uness otherwise sated co Dall Sabot Grom 2008 27880185 182 6.471 Tolover 0szI0 ~~ AGE—Ns 6ST aS —~C~« ZO 1 of eecor deal vate 2a __1 1 55% 10) The Cloud May Not Be That Incremental Nowadays, rarely @ conference call, presentation, or trade journal article goes by without some reference to the revolutionary benefits of cloud computing as extolled by the IT industy. However, what matters in the context of this report is what cloud computing ‘means for te IT industry from the viewpoint of investors rardware 28Carprrsuss™ In theory, if all workloads shifted to the public cloud, all this involves is the change in the purchaser of IT infrastructure from enterprise IT departments to service providers. For ‘example, we note that certain service providers such as Google actually manufacture their own servers, while other service providers such as Rackspace (covered by Credit Suisse ‘Small Cap Software analyst Greg Dunham) buy commodity components trom IT vendors such as Dell. The location and the ongoing operations of the underlying hardware and software are essentially transferred. When viewed from this perspective, the shift toward cloud computing will only be partaly incremental. As service providers grow in size and numbers, the efficiencies they gain will alow them to use the same amount of IT equipment (and headcount) io service a larger number of customers. Indeed, some 70% of respondents in the Credit Suisse IT Survey felt that, as a result of cloud computing, IT spending would be flat to dawn (as shown in F gure 30) Figure 30: Cloud Computing Will Reduce IT Spend in the Coming Years (Queston: Adoption of aud compu 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 100% ox + ot respondents ‘year from now Syearsfromnew Sears from now [Blincrease mDecrease EINot change ‘Source: Creat S680 IT Survey Febrary 207% From a vendor perspective, we believe thal the components of cloud computing are similar to existing datacenter architectures. Recently announced cloud offerings by all major vendors, including IBM, HP, and Dell are fundamentally similar to infrastructure ‘outsourcing services that they have provided through their services arms for several years The real change Is then to ensure that their portfolios are aligned with new purchasers—service providers, customers seeking private cloud implementations, and customers seeking public cloud services directly from the vendors themselves. In this context, the discussion about who is well positoned for the cloud is similar to much of this, report in comparing the strategic pros and cons of each company’s overall portfolio. For example, IBM has a broad portfolio from servers and storage to software and services, Whereas Dell predominantly serves the commodity part of the server market, which is valued by cloud service providers when building inrastructure for commodity services. The question then becomes whether the shift toward cloud computing could allow newer lenirants for IT infrastructure to win share, and we note that, according to the Credit Suisse IT Survey, Cisco is heavily favored to win incremental cloud infrastructure business, as 26Carprrsuss™ Figure $1: Cisco Is Likely to Be a Key Beneficiary of Cloud Deployments ‘ne scale of 15 7 unltely, 5" vor key how Rly ae the folowing vendors fo GAIN sigicant d HP EMe Microsoft Oracle Delt 250 270 290 310 330 350 370 3.90 Source: Gro Suis TF Survy, February 207%, arCarprrsuss™ Macro Trends Supportive of Tech ‘Much of our analysis deals with the microeconomics of the IT hardware sector in terms of the fundamental demand drivers of servers, storage, services, and PCs. However, for an Industry that is some $1.2rilion (Including hardware, software, and services) and that represents nearly 50% of U.S. invastment in private fixed assets, the macroeconomic fenvironment is a key consideration. After economic factors such as GDP growth, corporate health, CEO confidence, and several others are taken into account, we see a positive macroeconomic backdrop for an industry, which is ultimately the foundation for growth, efficiency, and productivity. As noted by John Maynard Keynes in Tho General Theory of Employment, Interest and Money, 1936: ‘there isthe instabilty due fo the characteristic of human nature that a large proportion of our postive activities depend on spontaneous optimism rather than mathematical ‘expectations, whether moral or hedonistic or economic. Most, probably, of our decisions to do something posiive, the full consequences of which will be drawn out over many days to ‘come, can only be taken as the result of animal spirits - a spontaneous urge to action rather than inaction, and not as the outcome of a weighted average of quantitative benefits ‘multiplied by quantitative probabilty" We would argue that such animal spirits are what lead to the cyclical volatility in investment, most recently Figureed by @ rapid dectine in spending in 2009 and subsequent snap back in 2010. The key issue now is where tech spending Is in relaton to the ‘economic cycle. To analyze this, we have relied upon three diferent approaches: frst, we Consulted our economic and equity strategists for broad indicators of actual investment in tech capex; second, we looked at specific levels of technolagy spending in the U.S. market (0.1% of net tech spending versus GDP}; and lastly, our own proprietary CIO survey. The overarching conclusion is that, despite the tech IT investment recovery in 2010, major economies are stil under consuming technology versus trend, and this will result in an upward bias to estimates in 2011. We highlight the folowing conclusions: ‘A conducive macroeconomic and corporate backdrop. A combination of strong balance Sheets, cash at record levels (in the U.S. at 6% of assets), robust FCF (at 4% of GDP. the highest levels since 1985), and high levels of business confidence mean that corporations hhave the continued firepower to invest for growth. Further, G4 business investment stands ‘16.4% of GDP, one of the lowest levels on record, yet tech accounts for @ whopping 4036+ of spending. The macroeconomic drivers behind corporate discretionary spending are very appealing and will prove to be supporive in the coming years for IT spending Tech is being under consumed. Having analyzed levels of technology, net investment rolatva to GDP, long-term rand, and dapreciation, we find that lavels of tech investment temain structurally below long-terms levels even after the recovery in 2070. For example, net tech investment (capex minus depreciation to GDP) normally runs at an average of 0.62%, with @ peak to trough of 1.4%; however, it currently languishes at only 0.1% This relate under consumption of tech in major economies adds to the argument that spending should continue to recover in 2011. In fact, net tech investment in 2008 for IT hardware was negate, implying that assets were depleted, and a recovery as seen ‘arlior in 2001 would suggest 2 0.3% recovery from current levels. Furthermore, forward Indicators such as CEO confidence, corporate profits, the ISM index, and durable goods orders imply positive momentum ahead. Cyclical recovery to continue, The end conclusion from a top-down perspective is that, despite an 8% recovery In rovenues in 2010, the levels of tech spending in 2010 provide a base for a continued cyclical rebound and revenue growth in 2011-12 that has an upward bias and will realistically remain above long-term trends until this under consumption is. corrected. Wo estimate the gap of tech investment versus recent tran levals is some 12% above current revenue levels. 28Carprrsuss™ “16 March 2011 Our proprtary Credit Suisse IT Survey is supportive of continued cyclical taiwinds. Based upon our survey of top IT decision makers at major global corporations, we find that, 163% of respondents expect IT spending to increase, with a consistent response across all regions. In aggregate, IT headcount growth is expected to expand about 6%, with spending increasing at significantly faster rates than in 2010, with the fastest growth in areas such as services, storage, mobile devices, and software. Corporations Have the Firepower to Invest Reassuringly, we believe that the overall macroeconomic backdrop for IT spending from corporations is positive, given several factors: Record levels of cash flow and a capacity to spend. As noted by our global eauity strategist Andrew Garthwaite, we are positive on corporate discretionary spending. There Js near-record underinvestment by the corporate sector in the G4 economies, and record highs in ree cash flow as a proportion of GDP, as shown in Figure 22 and Figure 99. With FCF at record highs, financing will be much easier, and there will be a strong incentive to Invest for future growth, Figure 52: G4 Investment % of GDP Figure 38: 64 Nonfinancial FOF % of GDP. Balance sheets have record cash levels. We note that there are very high levels of cash fon the balance sheet (6% of assets and 12% of market cap in the U.S. as of Q3 2010), as shown in Figure $4 and Figure 35. When balance sheets are this strong, It is often followed by increased M&A and fixed investment. Figure 34: U.S. Nonfinancial Corporates—Cash as a% of Figure 35: U.S. Nonfinancial Corporates—Cash as a % of Total Assots Market Capitalization ‘Source: Datosteor, Cedt Suse estmatee “Sours: Oates, Grek Susee estates Technology is tho single biggest category of U.S. nonresidential fixod equipment investment. An economic backdrop that is supportive to capex will directly benefit technology spending, owing to technology's majorty share of fixed asset spending. (See igure 96.) This share is consistently growing over time: as shown in Figure 37, net tech Investment (tech investment net of depreciation) has been growing at an 11% rate over the past 50 years: this compares with a 7% growth rate of overall nonresidential investment. Notice in Figure 37 that technology has resumed its preracassion upward tend and is likely to be the primary beneficiary of a continued growth in broader investment rardware 29Carprrsuss™ “16 March 2011 Figure 36: Composition of U.S. Nonresidential Investment Figure 37: U.S. Tech Spending as a % of Total in Equipment Nonresidential investment ‘Source: BEA, Groat Suisse estnaies Source: Datesoam, Credit Suisse ostinatos Business and CEO confidence is high. The CEO Confidence Survey (conducted by the Conference Board) depicted in =igure 38 has a strong predictive relationship with GOP {and interest rates. GDP is widely known to be coincidental with tech spending, and therefore this survey is @ good leading indicator of future tech spending. A total of 100, CEOs are involved in the quarterly report, and they are asked questions about their ‘economic and industry outlook. The measure now reads 62.3, which is above the historical average and bodes wall for fulure tech spending (a reading of more than 50 points reflects 1 generally positive outlook). Another closely followed technology indicator isthe NY Fed's Empire State Manufacturing Survey. The survey is sent on the first day of each month to the same pool of about 200 manufacturing executives in New York state, typically the president or CEO. As in the CEO Confidence Survey, they are polled for sentiment and ‘outook. A portion of this survey specifically asks for outlook on tech spending, and the results are depicted in Figure 39. This indicator also is above average and predicts & Continuance in tach spending growth. Figure 98: The Conference Board Measure of CEO Figure 39: NY Fed Empire State Tech Spending Survey Gonfidence (CEO Confidence Survey) ‘Source: The Conference Board, reat Suse estmate, “Sowoe: NY Fed, Great Suse satiate Orders indicators are calling for an uptick in production. One of the most useful leading indicators of production is the ISM New Orders index (Figure “0), an index inked to new orders by purchasing managers. An index score above 50 is considered positive, with levels above 60 being extremely positive. The past two readings have been above 60, indicating strong uptick in orders, which should be followed by a strong uptick in production and shipments. Durable goods orders (OGOs) tracked by the consus also appear strong (see Figure 41) and further support the positive incination of the ISM trend, rardware 30Carprrsuss™ “16 March 2011 the census number is helpful in that it is speciic to computers and electronic products Please note that DGOs are depicted via year-over-year growth rate, and that they are increasing, but ret as convincingly as they dd throughout 2010, Figure 40: ISM New Orders Figure 41: Durable Goods Orders—Computers and Electronics, YIV Growth Rate %. Levels of Tech Investment Are Below Trend Moving beyond the macro backdrop, we believe that levels of tech investment remain significantly below trend. As noted by our global equity strategist Andrew Garthwaite and our U.S. economist Jay Feldman, there is a net underinvestment in technology. U.S. tech investment appears low versus trend. Net tech investment as a percentage of GDP is at a record low, as shown in gure 42. The net subtracts out depreciation from {08s investment, We are below the 0.62% trend line of net U.S. technology investment to GDP. Many investment projects were put on hold during the recession, but the recapitalization that began in 2010 should continue into 2011 Figure 42: Net Technology Investment as % of U.S. GDP. THPGaaALELLAEE: [Net investment is approaching depletion levels. Looking at levels of net investment (capex. spending minus depreciation) of technology, we note that at the end of 2009, as depicted | net investment was nearing zero (capex and depreciation converging) CCleaty, the trends seen in the late 1890s may not retum, but levels of net investment near rardware aCarprrsuss™ zero can be interprated as an unsustainable deplation of assets. Capex spending beyond ‘depreciation levels incicates spending above replenishment, and in our view, is necessary for companies to use technology investment as a fuel for innovation and productiviy. The life of the install base is high, implying underinvestment. Another way to think of current levels of invesimant is the implication on the replacement Iife of assets. With net investment near depreciation, only one af two things can be happening: either assats have ‘a much longer useful life or there is significant underinvestment. While technology Innovation can certainly prolong the life of the installed base, wo believe itis unlikely to have stretched this far. We wil be back to prerecession technology capex levels within 12-18 months. Based on four estimates (rom our proprietary industry models on PCs, Servers, Storage, Services, and Printing) for 2010 IT hardware sector growth, which can be used as a proxy for capex grovth, a8 well as assumptions for run-rate depreciation, we have created a forecast for 2010-12 net tech spending versus GDP. This is depicted by the dashed line labeled timate in Figure 42. We assume 8% IT capex growth in 2011, and 6% in 2012 that will bring us to pre-recession 2007 levels of net technology investment as a percentage of oP. Underinvestment is pronounced in IT Hardware. To further examine the level of Lunderinvestment, we deconstructed Figure 42 (net tech investment as a % of GDP) into taste and sofware covpensnts, "save 23 Gpits cut US net hardware nvestment ona not software investment as @ percentage of GDP. The magnitude of underinvastment in hardware is clearly more pronounced and reinforces the emergence of a technology refresh cycle that began in late 2009, and in our view should continue into 2011 Figure 43: Net Investment in re and Softw: of US.GOP 09% on 05% 04% 03% 02% 04% WofUs. GDP Corporate profts are signaling pent-up tech demand. One of our favorite leading indicators Is corporate profits, a8 rising corporate profs typically result in accelerating investments in technology infrastructure. Nate in Fue <4 that corporate profits are currently growing at record levels, and we expect technology spending to directly beneft, especialy given that technology spending is below trend as a percentage of corporate profs. (See Figure 45). [As the corporate sector regains steady footing, and with bolstered confidence, it wil Ikely deploy its record cash piles into technology infrastructure. 22Carprrsuss™ Figure 44: U.S, Technology Fixed Investment and Figure 45: U.S. Technology Spending as a % of Corporate Corporate Profits, % Annual Growth Profits gga The Tech Pulse index is indicating healthy growth in the technology sector. The Tech Pulse index tracks economic activity in the U.S. technology sector (see Figure 46), but is limited in that i is @ coincident indicator. The index is constructed by the San Francisco eral Reserve from technology-specifc portions of five main economic indicators in femployment, investment, production, shipment, and consumption. In Figure 47, we depict the year-over-year percentage growth in the indicator, giving us a feel for continued positive growth, albeit less positive than 2010. This coincident indicator combined withthe Previously noted leading indicators (ISM, DGO, Empite State, CEO Confidence, and Corporate profits) contrbute to our generally positive outlook for 2011, as we have significant runway before any obviously negative signs appear on the horizon Fi YY Growth % Source: SF Fed, Cred Suisse estimates Implications for the IT hardware; the cycle continues, Based on the exhibits and data above, we can make a compelling argument for the relative levels of underinvestment in the technology sector and a supportive macro backerop both from ceincident and leading Indicators. When combined with our bottoms-up analysis, it helps us form several Important views with respect to 2011 and early 2012. As shown in Figure 47, on a year- over-year growin basis, several of our key end markets saw a cyclical recovery that started in mid-2009 and carried them through 2010. Based on the economic data and our bottoms-up analysis, we believe that growth will continue to remain resilient through 2011 and into early 2012. 33Carprrsuss™ “16 March 2011 Storage, PCs, Printers Figure 47: YI¥ % Growth Rate—Server ye y Our Credit Suisse IT Survey Supports This tthe end of January 2011, we conducted out semiannual Credit Sulsse IT Survey to help, Ls gain addtional insight into where we are in the IT spending cycle, Our survey was, completed by 70 top IT decision makers (Figure 48) Figure 48: Credit Suisse sur Fr Director 3% _ IT Manager se IT coneuitant | CTO CEOVProsident 2 7% % Source: Cred Suisse IT Survey ur survey respondents were positive on the economy and their IT budget growth into 12017. Nearly 90% described the economy as stable or improving, with 57% feeling that the ‘economy was improving! (See Figure 43.) More reassuring was the fact that 63% had an IT budget that was bigger than it was one year ago. (See Figure 50 rardware 34Carprrsuss™ Figure 49: Credit Suisse IT Survey—What Is Your Current View on the Economy? Figure 50: Credit Suisse IT Survey—How Does Your Current Outlook for Your Overall IT Budget Compare with One Year Ago? Respondents plan on spending more money in IT, and will grow IT headcount in 2011 Their positive views on the economy, along with larger budgets, wil Ikely result in spending more in IT products and services in 2011 than they did in 2010. (See igure 51.) This will come in the form of products, services, and increased headcount. (See Figure 52) Figure 61: Credit Suisse IT Survey—What Are Your Year-Over-Year Growth Expectations for Your IT Spend Next Year versus This Past Year? 210 0 20) me Figure 52: Credit Sulsse IT Survey—What Is Your Expected IT Headcount Growth Rate in 2011 versus 20102 What Was It in 2010 versus 2009 00% 10% 20% 30% 4o% 50% com ‘Sours: reat Suse 7 Survey, ‘Source: rea Suisse 17 Suvay. Services, storage, mobile devices, and software will be the biggest beneficiary of Incroased toch sponding. Consistent with our proprietary models, Fou reinforces our nesis that several areas of technology will continue to experience growth in 2011. This wil be led be services, storage, mobile devices, and software, Our companies with significant exposure to these businesses should experience the strangest growth, 38Carprrsuss™ “16 March 2011 Figure 53: Credit Suisse IT Survey—What Are Your Year-Over-Year Growth Expectations. for Your IT Spend in the Following Areas in 20117 Storage Sonteos Mobile Devees and Services PC Soourity Networking Seners 2.0% 1.0% = 0% 1.0% 0% 3.0% 4.0% ‘Source: Great Suse survey. rardware 36Carprrsuss™ Get Your Head in the Clouds It seems that rarely a conference call presentation, or trade Journal article goes by without some reference to the revolutionary benefits of cloud computing as extolled by the IT industy. However, what matters in the context of this report is what cloud computing ‘means for he IT industry specifically for investors. We address five key questions: What is cloud computing? Before even discussing the Impacts of cloud computing on the PC industry, its important to highlight that its very definition is loose and has become clouded, excuse the pun, in IT marketing hype. In its essence, we view cloud computing as the increasingly common choice among enterprises to forgo purchasing, owning, and servicing their own infrastructure and instead shifing some of their workloads to another service provider. The public cloud is an immature market in which new external service providers ranging from Google to Amazon (covered by Credit Suisse Entertainment Inteenet, and Cable DBS analyst Spencer Wang) to Salesforce.com (covered by Credit Suisse Software analyst Phil Winslow) effectvely provide some degree of computing power on pay-as-you go basis. A private cloud, in contrast, is essentially when a large enterprise centralizes its IT and acts as an internal service provider to ts business units. Why adopt cloud computing, why now? Any shi toward cloud computing is going to be ‘gradual; however, the fundamental driver is the desire to shitt from the reactive IT spending status quo in which 70% of IT spending goes to maintaining existing operations, whereas only 30% Is actually spent on innovation. The hope is that by outsourcing IT processes and inftastnicture, datacenter utilization wil increase and additonal resources will be available for business group to focus IT spend on innovation and gaining leverage from IT. Key drivers for increased adoption now versus anytime in the past are: (1} numerous reports of successful cloud computing implementations of key workloads Ike CRM, (2) the widespread adoption of vitualization of everything from desktops to servers to applications, and (3) the emergence of a slew of new service providers, including new offerings from traditional IT vendors, Private cloud computing has the added benefit of quantifying IT's contribution to the business, as it wil encourage the perception of IT as & service provider, and will slow for comparison with other extemal providers of similar services. Clearly, this is @ fundamental change in how IT is consumed and delivered by enterprises, an the vend is increasing on a long-term basis Why not cloud, the shit towards cloud computing will be slow. While this shit is clearly happening, we believe that the transition toward cloud computing will be gradual, as supported by our semiannual Credit Suisse IT Survey and discussions with industry leaders, Altnough the growth appears to be rapid on a small base, cloud expenditure is stl, {an insignificant portion of overall IT spending. A small subset of applications are cloud ready, and IT | slowly testing the viabilty of the public cloud delivery model for commoditzed, commonized applications and processes. IT's reluctance to go full steam ‘ahead centers around concerns about security reliability, and vendor lockin. There also is, fan element of self preservation involved, since outsourcing portions of IT can and will fesult in IT headcount reductions in areas that are commadiized and outsourced, This trend could take hold as public cloud offerings increase in diversity and are offered at cosi- effective price levels owing to the scale advantages afforded by pubic cloud operators Private clouds are of mare interest to enterprises (than public clouds) inthe near term, as they do not overhaul the way they are currently doing business and allows them to more efficiently leverage IT resources. How big is the cloud opportunity? Quanttying the cloud is fundamentally challenging given the fact that Gartner estimates the market at $68bn and IDC estimates it at $22bn for year-end 2010, Owing to the popular ofthe cloud computing trend and the marketing biitz by Microsoft, IBM, and Salesforce.com, almost all vendors are repositioning their existing offerings within the context of the cloud. This has made it dificult to determine what cloud computing realy is and is not. Significant portions of cloud computing revenue a7Carprrsuss™ fare existing products and services that have been recategorized to fuel the fad. Gartner {and IDC's projections for five years from now are $150bn and $55bn. This significant lispariy in forecasts is in itself very tellng and reflects a struggle to discretely define the bounds of cloud computing. Based upon our strict definition, we estimate the opportunity for public cloud service providers will be significant at $34bn in 2014 from $1 bn today. The opportunity will become significant aver multiple decades, and as such ithas attracted 2 nontraditional profile of service providers from the IT and telecommunications sectors, such as Amazon, Verizon, and Google. What doos it mean for IT spending; is adaltve? In theory, if all workloads shifted to the public cloud, all this involves is the change in the purchaser of IT infrastructure from enterprise IT departments to service providers. For example, we note that certain service providers such as Google actually manufacture their own servers, while other service providers such as Rackspace buy commodity components from IT vendors like Dell. The location and the ongoing operations of the underlying hardware and sofware are essentially transferred. When viewed from this perspective, the shit toward cloud computing will only be partially incremental. As service providers grow in size and numbers, the efficiencies they gain will allow them to use the same amount of IT equipment (and headcount) io service a larger number of customers. Indeed, some 70% of respondents in our IT Survey felt that, as a result of cloud computing, IT spending would be flat to down (as shown in Figure 20 below), Figure 5 Cloud Computing Wl Reduce IT Spendin the Coming Years 30.0% 70.0% ‘year from now Syearsfromnow ——_-S years from now [Blincrease mDecrease EINot change ‘Seure: Great Suse 17 Survey, February 207% What does it mean for IT vendors? We believe that essentially the components of cloud ‘computing are very similar to the existing architecture of existing data centers, While IBM HP, and Dall have jumped on the bandwagon in announcing cloud portfolios, our analysis suggests thatthe fundamental architecture is very similar. Even recently announced cloud offerings by all three vendors are fundamentally similar to infrastructure outsourcing services that they have provided through their services arms for several years. The real change is then to ensure that their portfolios are aligned with new purchasers: service providers, customers seeking private cloud, and customers seeking public cloud services directly rom the vendors themselves. In this context, wa would nots that the discussion ‘about who is well positioned for the cloud is similar to much of this report in comparing the strategic pros and cons of each company's overall portfolio. For example, IBM has a wide land strong portfolio from servers and storage to software and services, whereas Dell serves the commodity part of the server market, which is valued by cloud service providers when building infrastructure for commodity services. The question then becomes whether 38Carprrsuss™ the shift toward cloud computing could allow newer entrants for IT infrastructure to win share, and here we note that, according to our IT survey, Cisco is heavily favored to win cloud infrastructure business, given its broad range of services as well as strong networking offerings Figure 55: Cisco Is Likely to Be @ Key Beneficiary of Cloud Deployments ‘on scale of 7-5 (f= untkely, 5» very key, how bly are the folowing vendor lo GAIN sigsicant revenue 9s loud comoving adoption meresees? Cisco 18m HP EMC Microsoft Oracle 250 270 290 310 330 380 370 3.90 What is Cloud Computing? ur Definition of the Cloud Bofore discussing the opportunities and risks associated with cloud computing, itis worthwhile to be very clear about is very defintion. Indeed, while trade journals, industry veterans, and vendors increasingly use the term cloud or marketing hype, this can be Confusing for an investor. Ata very basic level, we belleve the adoption of cloud computing involves @ given CIO deciding that rather than owning and operating technology components that the time has come to move all or certain workloads to a service providers facilites. Cloud computing, at least in our view, then simply redefines how technology products and services are delivered by vendors and consumed by customers. \We expect enterprises increasingly to move away from building and operating their own computing infrastructure. Two good examples of this are: FedEx. In a Jan 24, 201, informationWeek article, a typical example of private cloud computing was discussed: "Just ast fal, FedEx opened a new data center in Colorado Springs based on this idea of general purpose computing. It uses commodity x86 servers, each with just a single 10-gig Ethernet cord into the back for networking replacing the bevy of wires of the past for host-bus adaptors, NIC cards, etc. Before applications move into the new data center, they are commonized—revised to use the same database and messaging technology, for example, so they can move easily among servers.” FedEx is using this cloud infrastructure inside is own data conter—a Prvate cloud—but Carter [its C10] says,"Workloads could easily shift to public clouds run by vendors such as Amazon and others, that made strategic sense down the road” 39Carprrsuss™ = The GSA (General Services Administration) announced in October of 2010 that federal, state, and local govemments would have access to cloud-based infrastructure ((aaS) offerings through a storefront at Apps.gov. This move allows vendors to aid ‘government with cloud storage, virual machines, and Web hosting services to ensure expansion of goverment IT cloud computing. Apps.gov offerings include on-demana self-service and resource pooling with nearly unlimited storage and automatic monitering of resource ullization. The GSA's laaS olferings also provide rapia slastcity and provisioning of virtual machines, storage, and bandwidth. Federal Crist Information Officer Vivek Kundra was quoted: “Offering laaS on Apps.gov makes sensa for the federal government and for the American people. Cloud computing services help to deliver on this Administration's commitment to provide better value for the American taxpayer by making government more effcient. Cloud solutions nat oniy help to lower the cost of government operations, they also drive innovation across ‘government To be clear, cloud computing is not a new technology; itis @ diferent consumption and delivery model for existing technology. It is riding the rapid uptake of virtualization (Giscussed later) and Web-based software in corporate IT envizonments. In many ways, it Continues the march toward outsourced IT and allows a company to focus on its core ‘competency (which typically is not IT). The technology elements comprising a cloud ‘computing solution have been in use since the inception of the Internet. However, the cloud computing terminology, conveniently refers to the go-to-market focus of several hardware, software, networking, and services vendors who are targeting consumers and businesses with outsourced technology and the provisioning of technology as a service. ‘The underlying goals are to hide the details of the technolagy from the end user, who can nen focus on achieving the benefts of the technology, and obviating the need for the technology expertise to install and maintain the technology. Public versus Private Cloud Cloud computing is divided into two broad categories The public cloud. The service provider offering a public cloud service manages all of the necessary hardware and software at their own data center. A common example of a public cloud service is Salesforce.com, which provides customer relationship management (CRM) sofware to businesses. A business is able to rent the software from Salesforce.com, and its users access the software from a web browser rather than buying and installing servers, storage, and networking. Per IDC and soveral industry analysts, cloud services are typically Accessible over the Internet via @ standard cont or browser 1 Standardized IT capabilites or services bu for a mass audience = Seltservice deployment and managemen ‘Elastic and scalable to accommodate growth '§ Incorporate usage-based pricing on a short-term or longc-term basis with ile upfront commitment © Hosted on shared infrastructure "= Transparent to the user with regard to the technical infrastructure underpinning the ‘There are three loosely defined segments of public cloud computing services. Public cloud service providers generally offer one or a combination of SaaS, laaS, and PaaS. SaaS is, by far the largest and most well known category of cloud computing services, and here are the general definitions: 40Carprrsuss™ = Saas (Sofware as a Service), where sofware applications are accessed via 8 web browser; = laaS (Infrastructure as a Service), which allows a business to rent basic computing services such as servers, storage, networking, and extra computing power lo augment or replace the purchase ofits own hardware and software; and = Paas (Platform as a Service), which offers development, testing, and deployment tools for writing cloud applications that are Web andior mobile accessible. The private cloud. In contrast, pivate cloud refers to on-premise technology. with access resvictes to a single enterprise, A private cloud allows the IT organization to act as the internal cloud provider to the internal users. Private cloud refers to the notion that big businesses would like their internal IT provided as a sot of services to various business Unis. Think of it as a business adopting public cloud concepts in order to build its own internal cloud, The delivery of IT as a service, however, is very compelling to executives at large businesses in that i allows them to quantiy the business value of discrote IT services; these intemal (T services can then be benchmarked against those of external providers. Invariably, this wil lead to potential internal political challenges, as business Lunt will be able to’ choose between the IT services provided by external vendors and those provided by internal IT. These concepts, again, are not new. Large IT organizations. Within enterprises have tried to provide technology as a set of services for a number of years, and traditionally have only been successful in doing this for afew discrete services Via the use of software only. The rapid adoption of virualization and operationsiautomation tools in the past five years is enabling the creation of internal resource pools of servers, storage, software, and networking: this allows IT to create intemal cloud services much like an extemal service provider would Be Careful of the Hype [As with any new technology, especially in a sales-intensive business, there is a risk that ‘marketing departments overtype the term cloud, as shown in Figure 55, Indeed, we note some recent observations by industry veterans. "The interesting thing about cloud computing is that we've redefined cloud computing to include everything that we already do. | can' think of anything that isnt cloud computing with all of these announcements. The computer industry is the only industry that is more ashion-drivon than women's fashion. Maybo Ii an idiot, but | have no idea what anyone Js taking about. What isi? I's complete gibberish. t's insane. When is ths idiocy going to Stop? Well make cloud computing announcements. I'm not going to fight this thing. But I don't understand what we would do differently in the light of cloud." Wall Street Joumal, Oracle CEO Larry Ellison, September 2008 "I don't know what it means. . 1 have nothing against the term cloud, my thought in the Industry is that when we talk about some opportunity, people tend to insert the word cloud, ‘and that’s going to be the answer to whatever question... (ike) i's magic.” ~ Mark Hurd, Oracle President and Former HP CEO, November 2010 To be clear, we believe thal there isa clear shift to cloud computing occurring, but equally we would not expect itto change the industry over nightCarprrsuss™ Figure 56: Hype Gyele for Cloud Computing, 2010 ‘A more than 10 years © before plateau ‘Soure: Borner, 2070, ‘Over the next few years, private clouds will be much more relevant fo the enterprise than public clouds. Owing in part to factors such az securlly and reliably concerns, large organizations are hesitant to use public cloud services beyond the few SaaS (hosted software) applications they are typically using today; they generally would rather keep their sensitive data on premise, Our IT survey respondents confirmed this notion (see Figure 57), which can be interprated as private cloud being the current priority, and public cloud adoption will see a gradual increase over time. Figure 57: External Cloud Infrastructure Has Room to Grow what prcont of your eal IT hardware sponding goesimight go o cloud inastucture vs. wekoad specie ang tte folowing pants in intemal cloud Extemal cloud Workload-specitic Infrastructure Infrastructure spending “Source: Great Suisse IT Survey, Pebrary 207% Given the need for many types of components to construct an internal cloud, several vendors have jumped on the opportunity to tout their abilty to be @ single sourcer or & majority sourcer of the technology elements. This obviously benefits the vendors by potentally increasing the size of their average sale and allows them to position several product ines at once in the course of a sales opportunity. Large vendors (HP, Del, Cisco, IBM, and EMC) will position their abiliy to provide an end-to-end solution, and smaller vendors (NetApp, Salesforce.com, CA) wil tout their ability to provide a hight 42Carprrsuss™ performance solution via a bestot-breed approach that ullizes their superiority in & specific portion of the solution. irtualization as a Driver of Cloud Computing Virtualization is part of how you do cloud, but tis not the cloud itself. There are many components that make up a cloud computing implementation, including servers, storage, networking, and software. As IDC notes, the software components enabling cloud computing include virtualization, automated provisioning, service-level. management, performance monitoring, consumption-based capacity optimization, and chargeback applications. A number of other management sofware products are also needed to enable ‘the aynamic resource scaling and automated provisioning capabiltias that are the halimark of cloud envitonments. We have intentionally included this long list of components to make it clear that cloud computing is not virtualization; the terms are not interchangeable (although one could mistake them to be, since they are often incorrectly sed in the press). However, virtualization is the Key enabler of cloud computing, and its rapid adoption is fueling the cloud computing conversation What is virtualization? Virwalization has been in use since the 1960s and was widely associated with |BM's mainframes. it allowed a large mainframe to be subdivided into Virtual machines—mutiple, separate, logical computing environments within the overall hardware computing environment. Currently, VMware, Citrix, Microsoft, (covered by Credit Suisse Software analyst Pnil Winslow) and RecHat provide virtualization software that can bbe deployed in today’s cliant-server architectures. VMware is by far the most successful Vitualzation software provider, and the introduction ofits virualization tools in the early 2000s sparked this new era in client-server computing. Virtualization software allows the partitioning of hardware resources inlo pools, a& opposed to the previous constraint of tethering speciic workloads to specific hardware devices. Virtualization allows for workloads to be moved seamlessly between physical devices, maximizing efficient use of hardware resources. A practical example of this concept would be the installation of Parallels or VMware software on @ Mac in order to run Microsoft Windows on the Mac Operating System, essentially creating a computer within a computer; imagine this same concept on large datacenter servers running hundreds and thousands of workloads. Different viral machines can run diferent operating systems and multiple applications on 2 single physical server, turning a single server into multiple servers. In addition to subdividing physical resources, vitualzation can be used to pool (or combine) physical resources lke storage; storage virlualzation can convert multiple storage devices into one large storage device. The same concepts also apply in application, desktop, and network vitualzation, Virtualization is @ very significant trend. This trend is 80 substantial, in fact, that large enterprise ClOs increasingly discuss data center inftastructure in terms of number of virtual machines rather than server-centrie meties. Chief decision makers at some of the largest IT shops confirmed this growing trend in our survey, as depicted in Figure 58 and 43Carprrsuss™ Figure 58: Virtualization of Servers Is a Growing Trend ‘uatzed ate folowng parts m tme? Figure 59: As Is the Number of Virtual Machines per Server (SPU atthe flowing pants tne? What are the benefits of virtualization? Companies traditionally dedicated a speciic application workload to a single specific machine since they aid net want applications to interfere with one another; this architecture prolferated with the advent of client-server architecture and servers that were cheap in comparison with mainframes, Over the years, this resulted in data centers full of underutiized machine: by most estimates, before Vitualzation, servers wore only 15-25% ullized. Virtualization software enables multiple workloads on the same machine, and hence allows for @ much more efficient use of hardware. Multiple servers can then be converted inte a computing pool that is exible and scalable. The results a significant reduction in: © Hardware, labor, and management costs—fewer machines can be purchased per workioad, and fewer people are required to manage the hardware infrastructure. |= Time—provisioning, maintenance, and testing of new workloads can be completed rapidly = Energy—when servers go unused, data center electricity and cooling costs are unnecessary high, wasting energy and natural resources. Software as a Service as a Driver of Cloud Computing—Cloud-Ready Workloads Software as a Service (SaaS) is by fr the biggest category ( 60) of the $11bn coud ‘computing market, as tis arguably the service that started the entire cloud trend. Saas's ‘most well known face is Salesforce.com, as it has been pushing the cloud computing concest for nearly ton years; iis fs expected to reach $2bn in 2011 revenue. SaaS has experienced steadily growing success in that it allows the enterprise to ty aut portions of cloud computing on a warkioad basis. As implementations grow in success and trust is gained, an organization can then ty outsourcing more workloads or infrastructure 44Carprrsuss™ Figure 60: SaaS Is By Far the Biggest Category of Cloud Computing ner estimate of 2047 Cloud Computing Compostion In Gartner's recent November 2010 SaaS forecast, they identified now well known categories of sofware are broadly being affected by the SaaS delivery model "© Saas continues to penetrate the CRM market, accounting for neatly 24% of total CRM. market revenue in 2009, SaaS in CRM exhibits more general market adoption, ranging between 11% and nearly 40% of total software revenue, depending on the CRM subsegment. SaaS may exceed 26% of CRM market total revenue in 2010. ©The content, communications, and collaboration (CCC) market continues to show the widest cispariy of SaaS revenue generation, with SaaS representing 4% of enterprise content management (ECM) and approximately 82% of Web conferencing, © Project and portfolio management (PPM) Is a fast-growing market for SaaS, with a compound anrual growth rate (CAGR) of more than 40% projected for the next five years. "= Office suites and aigital content creation (DCC) also continue to show rapid growth for SaaS, although starting from a smaller base, with a 31.7% CAGR and @ 36.7% CAGR, ‘Adoption is driven by new entrants in office suites but imted by broadband availabilty ‘and quality. = Revenue growth associated with SaaS will be double the total aggregated growth rates for both ERP and supply chain management (SCM), but adoption of SaaS within ERP and SCM vares based on process complexity. SaaS within ERP remains @ relatively small proportion of the overall market (in comparison with other software segments) 48Carprrsuss™ Despite these rapid growth exhibits, Software-as-a-Service is growing from a relatively small base and comprises less than 5% of the entre enterprise software market In SaaS overall is growing nearly 15.8% YY, at about 2.7x the rate of the broad software market's 5.8% rate. Most teling, however, is the magntude of SaaS revenve, representative of a stil fragmented market filed with many small players in the CCC (content, communications, and collaboratien) and CRM (customer relatonship management) software segments Figure 61: SaaS ls Less Than 5% of the Entire Software Market for the Foreseeable Future OTT Worldwide Sofware Revenue by Primary market 2009 __zo1o__zot1_zo1z__2013__2014 _CAGR(X) Hreprcaton YaH.606 159,309 198804 Ta, 68 196458 106,787 52% laspscaton development and desoyment eiars s72s0 71059 7sooa L773 Baees SK otal Enterorize Software Zrz 202 —Bea014 297 092 315,08 307. ta 301,814 50% orivide Soa SoRware ROveTI® By APBUCRTON 209-2078 [Sortware Market 2000 zoto__ zor 2012203204 CAG (X) fecc Zane —7a8 dart —a,1e 8,008 5905 196% loice Suites eat? aT 3% loce 97 ag a87% fort zore 284 ast 3288.78 129% lene Yee an eet Tato koe 400% lscw Gor 12 tO tarts tgs 13.7%, peu 70 anne Sanaa 0m. lother App Sorat oo 421428 sa 4780 nasa 14.0%, [Total Enterprise San Software 7935 ue to,ee2 tate 4928 1o.540 15.8% ‘Sowree: Garner (2070), IDC (2010), Credit Suisse estates To get a belter understanding of how key IT decision makers are thinking about their rmiterm purchasing plans for these specific software areas, we gave them several categories to prioriize with regard to SaaS migration, Not surprisingly, as Figure 62 depicts, they confirmed that CRM was the top priority, closely folowed by email, and then business inteligence and data storage. Notice, however, that computing cycies (generally faling in the laaS category) are last in priority, and security, ERP, and supply chain (all highly trusted applications) are also at the bottom of the prionty Ist. As IT organizations become comfortable with the performance, economics, and security of the small parts of clouds that they do use, they will accelerate their adoption of broader externalinternal cloud services, and this will play out in the growth of Inrastructure-as-a-Service (laa) and Platform-as-a-Sarvica (PaaS) as wall 46Carprrsuss™ “16 March 2011 Figure 62: Cloud-Ready Workloads (sea of tf 5, bomg not nly and 5 beg vry We) Source: Gre Suse IF Survy, February 2077 Why Adopt Cloud Computing? ‘There are several considerations that must be weighed in deciding to migrate specific workloads of portions of infrastructure tothe cloud delivery model. As with any investment docision in an enterprise, (1) economics, (2) risk, and (3) options should be key Considerations in weighing the total business value attributable to the investment. For the move toward cloud computing, Gartner found that these three elements and a few others, were top of mind in weighing the move toward cloud computing adoption. As Figure 63 depicts, Gartner identified the following six areas as being the most relevant: economics, agility, trust & risk, creativity & innovation, simplicity, and social impact rardware arCarprrsuss™ “16 March 2011 hed Against Six Key Factors lance weighing ebud co ‘reat and ‘Seure: Garter (Sepleriber 2008), Credit Suse estimates Economic and hard dollar benefits are not the only major sifferentitor in deploying applications in the cloud vs. (taditional) on-site. Innovation and agity are at least equally Important, as shown in Figure 64. Not surprisingly, however, trustrisk is the key area of ‘concem when discussing any outsourcing of IT—enterprises are reluctant to house their Sensitive data off-site, and the convenience of transparent IT is countered with the question: “where is my data, and who has access toi?" Figure 64: Creativity and Innovation are Top Benefits; Trust and Risk the Biggest oneern (na elativo scale. which coud feats provide highor vue than on-site ones? Economies Aalty Great 14 Simplicty Trust andRlek Sect impact ‘Searee! Garner Survey 2009), Crt Suisse eatmates rardware 48Carprrsuss™ Creativity & Innovation \when IT organizations are no longer concerned with running infrastructure they can focus fon solving business problems with innovative technology. Gartner, IDC, and several vendor studies agree that typical IT organizations invast over two-thirds of their time and budgets on day-to-day operations and in a reactive, tactical mindset as opposed to a proactive, strategic, “alignment with the business" way of operating, (See Figure 65.) This, is often owing to an inundation of requests to IT which are nonstrategic and breakifhc rolated. By definition, these low-value requests are often a poor allocation of a highly sklled IT worker's time. A cloud service provider can afford to time-multiplex these types (of requests among its shared IT staff, as itis centralizing talent for use among a broad range of client organizations and developing @ competency in commodity, mass IT requests, and services. When an individual business outsources these operational and Infrastructure services to @ cloud service provider, I rees up its internal staff to focus on strategic and innovative projects that can improve the core business. The vast number of available cloud applications allows an organization to experiment with new technologies with relatively litle cost commitment Figure 65: Cloud Computing’s Flexibility Offers Hope for Increased Proactive Innovation I egariaton peal pend ip 70% o ee buses on pert and manana jst ‘Source: nfomationiesk Anais Survey (2070) Economics, ‘The fist factor often discussed with cloud computing is the financial benefits associated with the pay-per-use model. Renting software and infrastructure (rom service providers) replaces the typically large upfront capex investment (associated with buying your own IT), with a predictable operations expense. Since cloud computing calls for the centralized management and automaton of technology as a pool of resources, the same amount of IT labor (headcount) can effectively manage a larger number of hardware and software components. This allows cloud service providers to leverage economies of scale and provision equivalent technology at a lower cost than most dedicated, internal IT shops. In theory, they should be able to pass these cost savings on to customers, and the majorly (of respondents to our IT survey feel that the cloud movement will reduce their IT costs going forward. (See Figure 66.) 49Carprrsuss™ Figure 66: T Decision Makers Expect Cloud Computing to Reduce IT S sad9pon of lous computing wi cause your over IT ‘Seure: Gre Suse 17 Survey, February 207% ‘The discrate areas of spending that organizations are looking to cloud computing for cost savings include: Lower total cost of ownership 1 Lower hardware and software costs Lower dataconter facility costs = Lowermaintonance costs = LoworT staff headcount [Athough one would think these savings are most prominent by leveraging the public cloud the private cloud model will enable the same intial cost savings to the overall corporation, as IT wil be forced into becoming an internal service provider tothe business. By acting as such, the IT department wil be forces to optimize its own infrastucture, as it risks being outsourced in some portion, f not completely (depending on the size of the organization). As Figure 67 depicts, organizations expect to reap the economic (and ancillary) benefits via private cloud fist, as they tly to squeeze more out of the Infrastructure investments they have already made. It can be argued that there is @ conflict Of interest here, as IT is vested in self-preservation. As cloud service providers ramp up Ccapabilties in discrete (T services, their advertised cost per service can be benchmarked against the total cost of providing a service via the internal IT department 60Carprrsuss™ Figure 67: Enterprises Expect to Reap Cloud Benefits with Private (Internal) Cloud First nd 10 ‘Sous: 0 no a ai ey ‘Sowcer IDG Survey (2010), Croat Susewostinales. SpeediAgilityiSimplicity Historically, when a business unit required a new workload or IT service, it submitted request to internal IT and it could be weeks of months for the approval, acquisition, and provisioning of the software and hardware necessary to put the workload inte producto, With the advent of public cloud service providers, desired applications are a few clicks away and can be sourced by an external service provider, potentially allowing business Units to disintermediate intemal IT. Clouds in general allow for the faster provisioning of technology, as the pools of resources are already in place to support customer expansion, In a January 24, 2011, InformationWeek article, Fedx's C10 provided thoughts on Cloud ‘Computing: “I started in the late "70s, right around '80, working this stuff, so Ive ridden every wave from mainframe to minicomputers to PCs to client server to object-oriented let's throw CASE in there somewhere back in the '80s—to Internet technologies. AS those waves crashed, we and everyone else have remnants of those things. What's happening nnow—for the first time, In my opinion—is there's truly a general-purpose computing environment that’s workload agnostic. You can throw different kinds of workloads on the ‘same computing server infrastructure, There's network convargence—all the natworks are IP, there's not a bunch of unique protocols. And there's converged storage technology. ‘Then there's software like Java that can make itll very portable across platforms.” temal cloud services allow organizations to leverage technology that previously was out fof their reach owing to either a lack of internal skilset or the abilty to afford the large upfront capital cost. Since external cloud service providers can pool talent in addition to the hardware and software resources, they will mora likely have a higher technical competence than most IT organizations. This benefits small and medium busine cannot afford the best level of talent or the hardwarelsoftware underpinnings. sos thatCarprrsuss™ Trust & Risk Security, service quality, and reliabilty are the top concer of enterprise cloud evaluators ‘These risks are often the mast cited reasons that enterprises will adopt private cloud over public cloud in the near term. Private cloud alleviates these concems, but note that iis harder to make an internal organization as accountable as an external service provider, as the external pravider can be replaced more easily. When IT is outsourced, the service provider can be monitored and managed to ensure delivery is suitable for the enterprise and that the delivery qualty is comparable forthe price paid Note in Figure 64 that trust and risk are the biggest detraction of off-site cloud computing, fs enterprises are not keen on storing critical data off-site, with concerns around both security and reliably. In addition to the safely of the data, a longstanding concer with any exterally hosted service isthe availabilly ofthe service and the ever present threat of losing the network connection, and hence losing access to the service, F small and medium businesses, however, i can be argued that a major cloud services Vendor is able to maintain a more secure environment owing tots abilty to leverage better talent and more reliable infrastructure. A recent MarketBridge study (quoted in CRN on Jan. 21, 2011) polling 1,000 North American small and midsized businesses found that security was @ top reason to mave applications to public clouds, as “48% of SMBs said hey believe data security would be better inthe cloud’ Social Impact Cloud computing is being positioned as the green solution, as it consolidates infrastructure resulting in a smaller environmental footprint through the lesser and more efficient use of hardware, energy, and real estate. Cloud computing is generally not geographically constrained, so IT can be doliverad in emergency situations and to developing nations, and hence faciltating greater global collaboration, Why Not Cloud? Secunty and reliability is of paramount concern with the public cloud. A recent IDC survey found, as depicted in Figure 68, that secury 's by far the most pressing concem around cloud computing. Organizations are not ready to gve up their most critical data, and this, was also highlighted in Figure 62, as our Credit Suisse IT Survey respondents strongly favored cloud for the applications that are most proven ta be cloud ready via years of Consistently positive feedback. New portions of IT will ikely be required to eam this same level of trust (via success stories and gradual adoption) before being fully rusted to @ cloud provider 62Carprrsuss™ Figure 68: Security Is the Overriding Challenge Facing Cloud Adoption not ata ety and 10 = ‘Source: IDC Survey (2030), Gre Suisse ostmalos The technology is immature. Given that cloud computing represents less than 5% of al IT spending, is relatively unproven versus technologies that enterprises have relied on for docades. CIOs are hesitant to rush into cloud computing at the rate that vendor marketing campaigns would lke you to think they are being adopted. The immaturity of the technology is highlighted as the second concern in Figure 68, Costs may actualy rise with increasing levels of cloud computing adoption. With private cloud computing, by defntion, companies must buy resources in pools and they must buy ‘excess capacity fo handle future workloads in order to mimic the elastic capacity of a true service provider. This may actually increase their cost of technology acquisition, hence partially explaining why the large IT hardware vendors are pushing the concept so aggressively (average customer order size wil be larger). With public cloud, there are several risk factors that can contribute to rising costs, among them are integration costs, customization costs, and vendor lock-in. Integrating a public cloud service with other portions of internal IT might pose challenges, owing to the use of disparate technologies. Since cloud services tend to be homogenous (one size fis all), they often must be Customized for an organization ‘0 fully leverage their benefits for @ specific use. As the customization level increases, the vendor lock-in likelihood wil rise. Vendor lock-in also becomes a serious concern when someone else is managing a mission ertcal portion of an enterprises IT infrastructure. Itis diffcut to diferentiate when using commodttized technology. By definition ofthe early public cloud, it is based on, and is the provisioning of, relatively commoditized technologies. As an enterprise, itis cifcult to differentiate your operation if you are using the same software, hardware, and IT processes as your competitor {who might also be Using the same cloud services). In the long run, it can be argued that cloud computing is detimental to creativity and innovation because it is the provisioning of commonized technology to the whole customer base, If IT is core to your business, it will be hard to differamtiate from competition if everyone is using the same applications and software 63Carprrsuss™ platforms. Therefore, in the near future, we expect that companies will only leverage the ‘most commoditized technologies via the cloud, and will continue to look to differentiate their operations by innovating internally. Large IT shops are clouds in and of themselves. In a recent CIO Magazine article Considering the use of private cloud versus public cloud, Intel's CIO, Diane Bryant, was quoted saying, “I have a very large infrastructure—I have 100,000 servers in production ‘and so | am a cloud. | have the economies of scale, | have the vitualization, | have the aglity. For me to go outside and pay fora cloud-based service—I can't make the total cost of ownership work." This is a classic response from large IT shops. Such IT departments, for example in the Fortune 500, control a level of IT expenditure that affords them leverage against vendors that is on a similar scale to whal many of the moder day cloue providers are touting. They are likely best served with consigering the private cloud approach, if not for its novel technology, then for the sake of transforming their approach to one of an internal service provider. Vendor viabilty is a major risk factor. A large percentage of SaaS providers are stil tolatvely small (eee service provider table in Figure 70) and their long-term viabilty has rot yet been established, If @ vendor goes out of business or is acquired, there is 8 downside risk of service aisruption or significant price increases, Who Are the Major Cloud Service Providers? [A discussion around major cloud service providers would mainly focus on public cloud service providers; in the next three to six months; however, we anticipate an acceleration fof external private cloud service providers. Although private cloud is primarily used to describe an intemal, on-site cloud deployment dedicated to an enterprise, we are baginning to see the emergence of externally hosted private clouds that are not generally accessible to the public; our IT hardware vendors will play a big part in external and internal private clouds, as will be discussed in the next section. The public cloud service provider industry is relatively fragmented, with Salesforce.com boing its most well known and largest constituent at $1.3 bilion in 2010 revenues. The provsioning of software as a service (SaaS) is most offen associated with cloud ‘computing, owing to the prevalence and popularty of SaaS vendors. SaaS allows enterprises to test cloud computing with relatively low-risk applications, and thelr growing trust is leading them to explore the hosting of infrastructure (laa) and the development of neir own custom applications (PaaS). Although some vendors only participate in a specific area of cloud computing, the top providers are increasingly blurring the lines. The top cloud service providers happen to be household names (Google, Microsoft, Amazon), and consumers who are using SaaS applications at home have come to expect the same types of applications in the workplace. Itis no surprise that our survey respondents (Figure 59) felt that several traditionally consumer names would become mare strategic in the Enterprise as cloud computing matures. 64Carprrsuss™ Figur 69: New Entrants Will Emerge inthe Enterprise Landscape matures? (clo of 1105, beng not ely and 5 bang vor Source: Gre Suse IF Survey, February 2071 ‘A recent joint sutvey by the UK Oracle User Group and Fujitsu (discussed in C10.com on January 25, 2011), found a wide discrepancy as to what a cloud service truly meant. Of ne respondents using cloud services, over one-half felt t was synonymous with SaaS, 21% felt it meant I8a8, and only 3% who were using cloud services associated it with PaaS, These percentages map well to the Gartner derived market estimates in Figure 76. Although Paa is the smallest segment of cloud computing, it is widely considered to have significant growth potential, owing to the proliferation of Intemnet-connected devices that can now access cloud infrastructure. The vast array of operating systems and hardware manufacturers fragmenting the client device industry, combined with the customer desire 10 have ubigutous applications, is driving the need for new software to be writen ang hosted in the cloud. Gartner follows about 60 companies in the PaaS market, the majority fof which are bolow $5 millon in total company revenue. OF the leading PaaS incumbents, Microsoft has the mast successful track record in the application development market Please see igure 70 for a sampling of prominent SaaS, laaS, and PaaS cloud providers 65carprrsuiss™ Figure 70: Public Cloud Service Providers “Source: Credit Suisse Eximates, 2077 Microsof, the largest software company in the world, is the largest looming threat to incumbent cloud computing providers, as its investing heavily in cloud marketing, product, development, and new data centers; CEO Steve Ballmer recently stated that ‘Microsoft is, botting our company on cloud computing.” Their strategy includes providing SaaS (Office 368 and Dynamics CRM), laaS (Azure), and PaaS (Azure); this further demonstrates that the ines between the three categories are bluring, especialy for large tech vendors. This, reinforces the notion that computing is becoming a utility, and there is increasing pressure to hide the technical details of now that utlily is delivered (much ike any other utlty you are familiar with). Additionally, Microsoft is @ well known name among enterprises. and already controls a significant amount of enterprise wallet share; their introduction of cloud services will provide for a smoother transition and reduced perception of risk. 66Carprrsuss™ ‘The use of SaaS in the enterprise is typically a tactical decision specific to a particular need, and is not a strategic outsourcing decision. Bul as more enterprises are ‘experiencing success with SaaS, they are more wiling to give laaS a try - opening the possibiliy for the cloud to capture an increasing share of their infrastructure. Amazon is considered the leading pure laaS vendor with services spanning elastic computing, cloud storage, and cloud content delivery. laaS typically requires a greater deal of trust since enterprise organizations are competent at running their own infrastructure. In contrast, many popular SaaS applications are in software categories that typically are not developed in-house, so the use of SaaS doesn't fundamentally change the function of the software, but rather its delivery. laaS involves locating the infrastructure offsite, and therefore new complications around security, network speed, service quality, and compliance must be considered; vendor reputation will be a key criterion in choosing a laa provider. In Gartner's recent SaaS study, they noted that an increasing number of enterprises are Using SaaS applications that were procured and deployed without participation from IT, creating both management and internal politcal issues. Thore aro hundreds of SaaS ‘companies to choose from, and Figure 7 below is a sampling of representative vendors in each major SaaS area, 67Carprrsuss™ Figure 71: SaaS by Enterprise Software Market, Representative Vendors fot som Sconce Esti ater hn sie Rape Mp aoa How Big Is the Cloud Opportunity? Before even artiving at any sensible estimate for the revenue opportunity for IT venders, we believe that itis important to arrve at the actual consumption demand from the user's perspective. There are three loosely defined segments of cloud computing services, with fone category—-SaaS—being the largest and most well known category of cloud computing = SaaS (software as a service), in which software applications are accessed via a web browser; = IaaS (infrastructure as a service), which allows 2 business to rent basic computing services such as servers, storage, networking, and extra computing power to augment €r replace the purchase of their own hardware and software; and 68Carprrsuss™ "= Pass (platform as a service), which offers development, testing, and deployment tools for wring cloud applications that are web andlor mobile accessible, Public cloud service providers generally offer one or a combination of SaaS, laaS, and PaaS. In such an immature market, ariving at any meaningful conclusion is challenging. Gartner publicly estimates the cloud services market at $68 billion dollars in 2010, and growing 20% YIY to nearly $150 billon to 2014, The breakdown of its forecast is depicted in Figure 72: Figure 72: Gartner's Estimate of Cloud Services, Worldwide, 2008-2014 ($8) Sarin’ Coud Forcast 3097010 7x 201220197014 CGR) 2000-14 ‘Business Process Services Avert 2 ee eT ee) war ‘Supply Management 4-18 a3 89 7 103 “ot Demand Management 28 a 400 556372 208 Operations tsar 203 26 3452388 aa ‘Business Process Services Total 5102579 a7.27 a2? 98811468, 78 Applications Total 579 785 9as tay 18312072 20 Application Infrastructure Plaorm irtuctre oO 01e 02 aks war Application Infastucture Total 015s aziz 030s ase a8 0998 446 ‘Systom infrastructure Compute Sereas 1921 ar 86 6 10s 564 Bacup Seroes 037 ast ss oars. 220 ‘ystoms infrastructure Total 170262 4a gt 92a 48 Cloud Services Total sag 6833 tod 12631488 205 ‘Sours: Gartner May 2070, In comparison, IDG's estimate in Figure 74 1s $22 bilion in 2010, and growing 25.5% YI to $55 billon by 2014. The fact that two major analysts have such a wide disparity in how ney characterize the size of the cloud market further reinforces the widespread belie that cloud computing isa loosely defined marketing term, which is yet to define a specific set of technology products and service, Gartner's numbers are inflated by what it defines as remotely provided business process services, including the orline advertising market that Google dominates. In order to get a better feel for what top IT decision makers felt is and is not cloud, we posed the question in Figure 73 and our respondents resoundingly felt that Gartner's inclusion of advertising does not belong within the cloud computing definition, 69Carprrsuss™ “16 March 2011 Figure 73: Respondents Do Not Feel that Advertising Should be Included in “Cloud” Insiing the coud computing market should we include the folowing? Cee ‘otrerpandentsaneweing Yee" ‘Seure: Gre Suse 17 Survey, February 207% Both the Gartner (Figure 72) and the IDC forecast (Figure 74) include services that existed long before the term cloud even existed. This is a key driver in our hesitation to fully buy Into the accretive impact of cloud computing. A good portion of services classified as cloud ‘computing will come as no surprise to our IT hardware vendors, as they are participating in those sales opportunites alraady. Figure 74: Worldwide Public IT Cloud Services Revenue by Sagment, 2009-2014 (SM) To0a Share 2000-2014 F044 Share zoo 2010 zor gota ata) CGR) epreatons B18 Yast 3040 Ye382 17470 20580 40.7 204% 37.7 Anpicaten davelopmant 1647 «2264 =«9.190 425 G.07S.SRSIR 190 39.2% 8S and deployment ‘Syston inastrn ass 4a81 sare 704 ar? thats mas arate as tere 2958 ano 460m 7B Stk Storagotbase) 142 2140-2908 4.098 Sate 7968 8638133 Total sss 22,73 2B T3A 5918S 55457000 2T4% 100.0 rome Si% 200% 25% 22-10% 26.50% Source: 106 (Apri 2010 In the Credit Suisse IT Survey, we asked 60 top IT decision makers which of the two forecasts they relied upon, and the results are reflected in Figuro 75. Over 50% of our respondents felt that neither forecast was the one on which to count rardware 60Carprrsuss™ “16 March 2011 Figure 75: The Majority of Respondents Do Not Rely on IDC or Gartner Cloud Forecasts {Garner pubily estimates the cows serices mart at S6BBilion in 2010, ard $150 Blan a 2014. companson, DC's sstmato 5 $22 Billo in 2010, and 85 Bien by 2074 Whose forecasts rah? “Source: Gret Suisse IT Survey, Pebrary 207% ‘The majonty of vendors that are considered to be cloud providers do nat disclose specifics for their cloud-related revenue, so there is @ tendency to take pre-existing revenue streams and categorize them as cloud, Given what our surveys are telling us, we prefer the more conservative cloud estimate using Gartner's numbers and strpping out pre- existing business process services. We believe that the market is best approximated by taking a subset of Gartner's Figure 72 numbers, with the following mapping: SaaS Applications, PaaS = Applications Infrastructure and laaS = Systems Infrastructure, Figure 76 below is @ subset of F'gure 72. The numbers in Figure 76 represent the undoubtedly new products snd services comprising accretive IT spend, Figure 76: WW Public Cloud Services for SaaS, laaS, and PaaS (SM) 2009-2010 «201120122019 2044 2009 share ZO00TO14 201A share CAGR ‘Seas (ware sea Sones) 8790 7580 9Bu0 TZaTO_jesIO aay TeTOR 79.00% 60 TOR JaaS (nastuctre as Series) 1700 2820 «4.400 «6 600940012400 22.20% a8aOH 36.30% PaaS Platomara Servos) 168212805] NOK AGO 200K Tota! Tea 1038214595 segt026aT 34.078 mo% ‘Sours: Gant, May 2010 Is It Incremental to IT Spending? The trend can't be questioned, but the rate of growth must be questioned. Our server, storage, and services models give us insight info how cloud computing will influence Infrastructure demand and the numbers are simply not there right now. The adoption of private clouds is an architectural decision that fundamentally doesn't alter the type of technology components Enterprise IT will purchase, 6 the components are already fundamental to running any data center. When an organization decides to outsource technology to a cloud service provider, they wil direcly spend less on infrastructure equipment. but the purchase of infrastructure is still made by the cloud provider rardware 6Carprrsuss™ Organizations expect to save money as they adopt cloud computing in the coming years. (See Figure 66.) This might be good news for customers, but poses a challenge to our IT hardware vendors, as IT hardware purchasing could be increasingly consolidated to 9 smaller number of ever-growing service providers. Instead of each customer buying their own IT hardware and software, a larger percentage of customers wil get thelr IT indirectly through cloud services from companies like Amazon, Google, and Rackspace. This will likely be counterbalanced by the large IT hardware companies providing their own cloud services with their own equipment as a foundation igure 77, however, clearly depicts a key diferentiaor between public and private clouds. ‘Server unit shipment growth and revenue growth within private clouds are significantly higher than in public clouds. Daspite a rapid unit growth in public cloud servers, the types of servers are indicative of the public services themselves—relatively commodiizes, ‘one-size-fts-all technology applications. The server ASP in public clouds ‘= significantly below their private cloud counterparts. This leads us to believe that mega infrastructure technology providers will target the private cloud approach in the near term. Figure 77: Worldwide Public and Private Cloud Computing Server Revenue and Shipments, 2009-2014 AGRI Revenae(iMy Pusie Ce er ‘Shipments (000) Purse ee Pave wiser 2a 30k Our covered companies will supply components to the public cloud providers, and will reposition their products within the context of private cloud, This wil likely net sum to zero, {as the cloud providers will buy the software that IT taditonally would have purchased. Our analysis focuses on identifying an unforeseen disruption to our IT hardware companies, {and we simply do not foresee 8 discontinuation of business as usual; this is @ revenue shit story and not an accretive revenue story. A key example would be Dell, which is widely known to supply hardware to large cloud computing companies ike Rackspace (RAX-US}; despite this fact, Del's company performance has not shown any fundamental shift as it continues to do business as usual {As shown in Figure 78, top IT decision makers told us three things about cloud computing (1) it will shit the balance of power among IT vendors, (2) they expect it to save them money (fainforcing Figure 30), and (3) it represents @ major paradigm shit. The key beneficiary of cloud computing will be the existing SaaS vendors. As evidenced in Figure SaaS Is the bulk of cloud spending right now, as organizations are trying the public Cloud on @ piecemeal basis, This is good for SaaS vendors as they should take share from ‘existing clientiserver based software vendors, As for our covered IT hardware companias, they will need to expand their software offerings, elther intemally or via acauistton, to leverage the oncoming cloud computing rend. 62Carprrsuss™ Figure 78: The Expectations for Cloud Computing Are High ‘5 Strongly ree. Cloud comoving, 8285, 1a08, ard PaaS. ae a “Source: Gret Suisse IT Survey, Pebrary 207% Who Is Positioned Well for the Cloud? Owing to the level of press, customer inquiry, and analyst coverage devoted to cloud ‘computing, the majority of technology hardware and software vendors are positioning their products within the context of cloud computing. Regardless of wnere a cloud is located ‘extemal of inlernal—it requires a scalable, integrated architecture composed of server, networking, and storage hardware combined with application and operations sofware. Large technology vendors who have a broad pontfelio of technology elements have ‘embraced the cloud concept because it is @ logical argument as to why @ potential Customer should buy more tachnology al one time (in an atlempt to increase average sale size). IBM has the longest history in sellng these integrated stacks of technology under different names, such as converged infrastructure and adaptive enterprise ‘Along with Cisco and Microsoft, IBM is leading the marketing push for cloud computing. In igure 79, that IBM was voted the most strategically positioned among our covered IT hardware companies. VMware, with its leacing virtualization products, isthe key enabling technology for cloud computing, and hence its strength among our respondents, Microsoft is well on the path to providing public cloud services, and as a software company, stands to gain considerably from the adoption of SaaS. Cisco, as the dominant provider of networking infrastructure, is viewed by our respondents as being the most crucial hardware company tothe proliferation of cloud computing, as the cloud will depend on the integrity ofthe network for its delivery, security, and performance. 63Carprrsuss™ Figure 79: Sofware and Networking Are Strategically Positioned for the Cloud ow wile adoption of cout compuing ate! the sratgicinportane fe. wale! share) of the fotowing venaore? opie ‘Soure: Grect Suse 17 Survey, February 207% Technology vendors with only a portion of the entire stack have long pushed the bastof-breed argument with the reasoning being that small, specialized technology providers have the best performing technology in their respective area, Within the context (of providing a comprenensive technology solution, the best-of-breed approach argues that the customer is best served to choose the best vendor in each area and integrate multiple technologies for a peak performing solution. This has been a recurring theme in IT for decades—an almost cyclical question—as to whether technology solutions are optimally purchased as systems or chosen piacereal on the basis of component merits. We gathered feedback from dozens of CIO's in our Credit Suisse IT Survey, and the response was resoundingly in favor of the best-of-breed ‘approach (Figure 60), This is not to say that the mega-vendors are at a disadvantage in private cloud deploymants since they also sell their solutions on a component basis. What is really tells us, however, is that customers are hesitant to buy intact private cloud ‘computing solutions such as the cloudin-a-box touted by Oracle. Vendors who sell a system-in-a-box often face an uphill adoption challenge due to the fact that the majority of, customer IT architectures are heterogeneous and consist of several vendor technologies. Self contained systems, therefore, can only address a portion of a customer's business challenges, and their performance advantage is isolated to specific, often high-end applications. Furthermore, the lack of in-house expertise required to implement and operate a proprietary, high-ond systemvin-a-box may require the customer to take on Considerable time and labar expense (consulting services). In contrast, modular systems leverage pre-existing skilsets and technology, and customers seam to be leaning toward more modular, best-of-breed solutions Figure 80 for thar private cloud initiatives. 64carorrsuss™ tomva sor Figure 80: IF Decision Makers t-of-Breed for Private Clouds. yo woe to build for euro have) an iloral private lu, you wou “Sour: Great Susee 17 Survey, Pebrary 207% Regardless of whether an IT hardware vendor (lke IBM, HP, Dell) is pushing a best-of-breed or total solution approach, there are fundamentally two groups of buyers, they are targeting in the context of the cloud concept. (1) The same group of IT buyers and organizations they've called on for years and \who are interested in private cloud infrastructure and services (and potentially public cloud services from the IT vendors themselves), and (2) public cloud service providers who are potentially new buyers of IT infrastucture The private cloud is typically aimed at traditional, internal IT buyers, and IT hardware vendors are pushing their pre-existing technologies to these buyers within the private cloud architecture. All service providers offering public cloud services, whether they aro PaaS, laaS, or SaaS require data conters and integrated technology stacks. Therefore, ‘most of our hardware technology vendors are targeting data centers in one way or another, ‘and their cloud offerings are sultabe for the public or the private cloud, ‘An increasing trend we expect in the coming months is a slew of IT hardware venders tentering the cloud service provider space with both public and private cloud offerings. HP. and IBM have offered hosting and managed services for several years, but they are not publicly advertised as to not confict with thei hostinglcloud provider customers and other Channels which buy theit infrastructure in bulk. Our convictions were reaffirmed by the Credit Suisse IT Survey. in which our respondents, in Figure 61, indicated that there is a strong likelhood that Microsoft, IBM, Cisco, HP, Oracle, and EMC would become public cloud service providers in the near torm to midterm, rardware 65Carprrsuss™ Figure 81: Large IT Vendors Will Soon Be Public Clous Providers Inthe next 12:24 months. wich ofthe folowing vendors wil ely Become public cloud service providers ~ Ess = — sa “Source: Gret Suisse IT Survey, Pebrary 207% In a recent interview with Microsof’s Windows Azure team, Dell cloud evangelist Barton George said he foresees the distincton between private clouds, where Dall currently plays, and public clouds dissolving In much the same way that we really focused on distinctions between Internet, intranet, ‘and extranet inthe early days of those technologies, there is perhaps an artificial level of distinction between virtualization, private cloud, and public cloud. As we move forward, those diferences aro going to melt away, toa largo extont. That doosn't moan that were rot going to stil have private cloud or public cloud, but we wil thnk of them as less distinct from one another. I's similar to the way that today, we Keep certain things inside our Irewalls on the Internet, but we dont make a huge deal of itor regard those resources Inside or outside as being all that distinct from each other. 18m IBM performed very wel in the Credit Suisse IT Survey: among our covered IT hardware Companies, twas fist in strategic positioning (Figure 79), frst in likely revenue gain from cloud computing (Figure 85), frst in likelihood to become a public cloud vendor (Figure 8), second in servers for private cloud (Figure 83), and fourth in storage for private cloud (Figure 84). IBM has a vast array of cloud computing hardware, software, and services. No ‘company in the word can match its depth in hardware, software, and services: although HP has offerings in each of these three areas, IBM's services and sofware organizations are significantly largar than what HP can bring to bear. IBM's cloud product portfolio is, extremely broad and premium priced when compared with competion. Some example include: application software available on Amazon Web Services (AWS); CloudBurst, a private cloud appliance that competes with the Vblock solution from the VCE colton and Exalogie solution from Oracle; the IBM Smart Business Test and Development service nat runs on IBM Cloud; SaaS integration tools like Cast Iron; IBM Application Development Services for Cloud; Rational and Tivoli software for cloud; IBM 68Carprrsuss™ Implementation services for cloud; IBM Federal Commurily Cloud: IBM Smart Analytics Cloud and 18M Cloud Labs, to cite many examples. Also, IBM acquited Cast Iron in April 2010 in order to ease customer transition to public and private cloud architectures, Of the three major infrastructure vendors, IBM has the strongest potential to organically offer, converged SaaS, laaS, and PaaS services. HP and Dell would raquire partnerships to offer PaaS, as they do not have IBM's proficiency in application development and deployment. On January 27, 2011, IBM announced new partnerships around LotusLive public cloud services for providing email, Web conferencing, and collaboration. This is largely a counter to Microsoft and Google's momentum in similar cloud based tools; it also ‘mentioned new initiatives with SugarCRM (salesforce.com competitor) and Ariba Figure 82: No One IT Supplier to Handle It All Inthe year 2016, you could choose to gurchose the majorly of your Thardware/eotware/eences rom ‘ne vendor no woul “Source: Groat Suisse IT Survey, Pebranry 207% HP HP offers cloud service under many names, depending on the organization spear-heading the customer effort. Traditionally, however, its cloud offering was dual pronged: the Converged Infrastructure product offering and the services led CloudStart offering Converged Infastucture refers to a modular architecture consisting of various customizable products from HP's software, storage, networking, and server offerings Cloud Start leverages Converged Invrastructure products to enable customers rapidly to build private clouds. This gives HP flexibility in capturing customer business based on the customer's preference for the do-I-yourself or consuling-led approach. These offerings. are primarily aimed at private cloud installations, but could easily be positioned to public cloud providers as well. HP's server hardware was a favorte response among our survey participants, as shown in Figuro 83 orCarprrsuss™ Figure 83: HP Servers Are Well Positioned for the Private Cloud Wien sorer vendors a bet postioned to gain share as» result of your r-arhiecting for lea loud? HP eM Cisco Ucs Delt SunlOracle Fujtsu Whitebox x86 other 090 4.90 = 200-300 00S 5000 ‘Source: Crect S650 IT Survey, Febrary 207% (On January 25, 2011, HP announced a cloud compute service for enterprises called "HP. terprise Cloud Services-Compute;” this is effectively an enterprise private cloud hosted ‘at HP data centers. It also announced an integrated package of hardware, software, and services for building private, internal clouds under the name CloudService. The bulk of these offerings leverage HP's existing product and services portfolio and leverage HP's breadth to compete against bestof-breed solutions ftom Oracle, Cisco, and EMC. We ‘expect similar upcoming announcements from Dell and IBM HP's vast portfolio of haraware, software, and services products is only matched by IBM, and its breadth gives them extreme flexbiliy for the many faces of cloud computing whether private, public, or a combination of both. Ithas the ability to push the entire private cloud solution, best-of-breed components, or postion itself as a public provider. EMc ‘Owing to its 80% stake in VMware, EMC is undoubtedly strategically positioned in cloud deployments, as depicted in Figure 7°. Virually (no pun intended) all cloud deployments ‘must begin and end with vitualization, and this fact places EMC with an entry point into & majority of cloud opportunities, VMware controls the foundation of cloud computng— Vietualzation—with over a 50% market share according to IDC. EMC positions the Atmos brand as iis cloud storage platform. Atmos is a combination of software and storage hardware that is optimized for global, mutitenant, scalable pools of storage that can be leveraged within @ cloud deployment. The Credit Suisse IT Survey results depicted (Figure 84) EMC as a customer favorite in storage positioning for intemal (private) cloud deployments 68Carprrsuss™ Figure 84: EMC Storage Is Well Positioned for the Private Cloud ous? eMe NetApp HP eM Delt Hitachi SunvOracle Othe 000 1.00 © 200 -3.00 400 S00 6.00 7.00.00 ‘Soure: Grect Suse 17 Survey, February 207% [though EMC alone appears to be a one dimensional storage vendor, in November 2009, itentered into a joint venture with C'sco and VMware named VCE (The Viral Computing Environment Company). The VCE coalition is tasked with expanding customer adoption of he three companies combined cloud solution, named Vblock. This combined solution is a very potent competitor to IBM and HP comprehensive technology offerings. They go to market by enabling a community of systems integrators, service providers, channel partners, and independent software vendors (ISVs). The coaltion has also established Untied presales, professional services and support capabilties to simplify customer fengagement. This best-of-breed solution is well positioned to capitalize on the customer preference for using best in class products to build out their private clouds. (See Figure 80.) Del Unlike HP and IBM, Doll lacks essential storage, networking, and software components to provide an end-to-end cloud offering. Dell's cloud strategy Is loosely defined around ‘existing components and services, which can provide elements of a cloud solution. Their rocont acquisition of Boomi, a SaaS integration company indicates a desire to be a more strategic cloud provider: Boomi faciitates the connection of private and public cloud software applications (similar to IBM's Cast Iron acquisition), Dell, nowever, ranks last among our covered IT hardware companies in strategic positioning (=igure 79) and potential revenue gain (Figure 85) for cloud computing. This is despite the fact that Dell is a commonly mentioned supplier to several of the top cloud providers, But, as we've discussed earlier, they are providing relatively commodiized hardware within a (cloud) paradigm that is focused on squeezing cost out of IT. it seems logical for Del to instead consider providing its own cloud services based on its hardware products. A limitation for them, however, willbe thelr lack of depth (compared with IBM {and HP) in sofware, Dall does have a growing presence in IT (operations) management software, and ths is a Key pisce to botn Becoming and supplying a cloud provider During the last week of January 201, rumors began circulating of Del's potential entry into the public cloud arena, widely viewed 8s @ me-too offering alongside HP and IBM announcements of new cloud services. As a major supplier to several cloud providers, Dell, will have to carefully balance the upside to these services with channel conflict it may incur by competing with service provider customers. If Dell were to offer laaS andior PaaS. services, it would likely leverage its strong partnership with Microsoft and base the offerings around Microsot's Azure products. 69Carprrsuss™ NetApp [NetApp is the second.targest independent Network-Attached-Storage vendor in the world behind EMC and No, 3 to EMC and IBM for extemal disk storage. Storage s a key element of cloud architecture, making NetApp invariably part of the cloud conversation, They acquired Akorri Networks in January 2011—a vitualization management software ‘company-—in a mave that gives them access to Akart' large viralization customer base as @ means to cross-sell NetApp storage devices. NetApp is looking to leverage vitualzation in a manner simiar to EMC, giving it access to storage opportunities in the private and public cloud. As shown in Figure 84, our IT survey was very favorable to [NetApp’s storage potential for cloud computing ~ it was second only to EMC. [As a best-of-breed storage vendor, NetApp has strong potential to play to customer private cloud preferences for piecing together the best components for a high performance solution (25 our survey reinforce in F-gure 60), Like EMC, NetApp has a very strong relatonship with Cisco and VMware, called the “Imagine Virually Anything” initiative, allowing it to compete with HP and IBM for storage hardware and storage software portions of cloue deployments Cisco Despite not being one of our covered IT hardware companies, we must mention Cisco since they are one of the main beneficiaries of cloud computing. In fact, our survey respondents resoundingly mentioned them as the company with the most to gain strategically (F'guro 79) and on a ravenue basis (Figure 85). Note in Figure 79 that iis no coincidence that the largest virtualization, networking, and software companies are at the .0p ofthe lst. Each ofthese three areas are a mandatory element of cloud computing, ana fone would expect that cloud computing wil bolster their strategic importance. With Cisco's recont entry (the UCS products) into the servar space, its position Is arguably stronger and should make them a formidable competitor for HP and IBM forthe entie data center isco UCS (Unified Computing System) is @ highly integrated network, server, and software offering that sarves as a private cloud “starter kt’, and is a direct competitor to Oracle's self-contained Exalogic product. A customer evaluating Cisco's UCS products is generally not looking to only buy servers, but instead is looking for a comprehensive data Center solution that includes servers, networking, management tools, and storage, At the core of Cisco's UCS architecture isthe tight integration of internally developed networking land server bulding blocks that leverage Cisco's significant IP and innovation in networking {and /0, This combination of products is aimed at competing with IBM and HP, both who have comprehensive data center offerings. Owing to increased competition from IBM and HP for the data center, Cisco felt it was necessary to enter the server market as it was losing leverage by nat directly contolling the server IP. To round out their cloud offering Cisco will often partner with EMC or NetApp for storage and VMware for vieualization software 0Carprrsuss™ Figure 85: Cisco Is Heavily Favored to Gain Revenue as Cloud Computing Grows Cisco oracle Microsoft me He Del 2 22 24 26 28 2 32 34 36 a8 4 ‘Soure: Grect Suse 17 Survey, February 207% rac [As one of the largest enterprise software companies in the world, Oracle controls a key flement of the data conter: middleware and database software. Software Is the driving force behind the majorty of cloud computing as it exists today, given the dominant share. held by SaaS implementations. Our survey respondents (F'gur® @5) feel Oracle stands to gain significant revenue from cloud computing, especially as it migrates its applicatens to public and private cloud delivery models. Oracle's January 2010 acquistion of Sun gave them a server and storage offering that makes them a stronger competitor to HP and IBM Oracle's most well known cloud offering Is thelr recently (8/2010) announced Exalogic Elastic Cloud. Ths is a self-contained cloud-in-a-box system containing 30 servers, each loaded with ‘wo sixcore processors for @ total of 360 processor cores. They are interconnected with each other and storage via Infniband connections, Oracle also touts the Exadata product line, which is often thought of as the “Oracle Sun Database Machine", as it combines Sun server and storage technology into an integrated system that is optimized for running Oracle's marketleading database sofware. This architecture is aimed at reducing total cost of ownership, and Oracle has publicly referred to Exalogic as a coud.in-a-box."Carprrsuss™ “16 March 2011 PCs Disruption Coming (Over the past five years, PC industry unis and revenues have grown 12%/4%, reaching 368mn units with revenues of $248bn in 2010 and representing one of the largest segments in the IT sector. Throughout this period the PC evolved to higher performance specifications, has become more mobile (with tablets boing the newast addition) and has Increased its penetration in emerging markets. Our extensive proprietary analysis of the PC industry leads us to 10 important conclusions over the next five years. Figure 86: Tap 10 PC Conclusions J Werhave Dui he Fok econometic oda Tar PCR [oabed Gh 2 COUNTIES and 1 DODY Ga POG), Ad Can SISIERTy Fave 8 BC relatonsip between PC avorstilty and PC panetaton pr eapin gen R? values for our regression Detwoan 71% and 8% 2 Bases on our propria Bil of Matrais analysis (BO), wo Sosove tal long form an accopiabio (ual) PC! tad ean sella ah varege past of 8200/8800 respectively 3 Based on our expectalors for GOP growth’ ASP decinas ol 5-84 long-term we expect consumer un growth of 18%/18% 2014/12 (392% extablets) and STH UT (ir ox'ablts) a Aste av af eammercal PCs docires fam 6 years last eat fo S Years over tha nex wo ears we: {asesein 2011112 (13%!10% extablats). Lang orm we Salave hetessng eommercal PC panetaon of growin of 119 (8% oxsablot) '5 Combing our consumer ane comercial PC model, wo oxpec overall @lbal) PC uit grow oF 18% 7 n ZOTIF2 BHIBK ox tabi) and 14% U7 (3 ex aba). & — Weexpect emerging make! unt growl of 19% long erm, driving unite om these regions fo 5% a lol PC demand by 2015 from ct unt wont of 469 toy. : 7 Tablets wil 9 dominant PC catagory veraus nelbeaks given media aptinizato,o seat up fundtnaly, i) oplinized mabie O8 iv) arse hannels ares) instanton sandby, : {8 Tho vansson to abet wil afore PC vendors a ones ina etme opportuni o break the Wintel duopoly and capture higher valve sha {8 Based on our expectation fra shifin volumes down PC pie ers, we expect lad perealon lobe highest In lowe ts glen De ampite powor needs those aves, Long erm wo estimate abou wil penetrate **% ofthe Sesilop matkat and 38% ofthe rctiona fabie PC market resultagn uns f 208m and revere a $12¢0n {0 We expect radtonal PC volumes (te martet exuding tablet) to raw 3H LT and docine 6 in rovenue arma ‘Source: Company data, Groat Suisse estimates. 1) An econometric approach to modeling the consumer, a highly elastic market. We have developed what we believe is the frst econometric model for consumer PC demand using {2 cross section across 42 countries and based upon over 1,000 data points. We find that there is a statsticaly significant relationship between PC affordability and the PC penetration per capita, (with R squared ranging between 71% and 86% for our multiple regressions), Based on these fundamental relationships, we demonstrate thatthe elasticity lof demand remains above 1.0 and this means that a move to lower price points will drive Incremental volume. 2) POs at $200 are not that far away. Based on an extrapolation of product teardowns, for Wich we discuss two devices which have sold in significant volume: Del's Inspiron Mini 9 {netbook}, and the Apple (Pad (tablet). We look at how the BOM may evoe in a Competitive market and demonstrate that an average quality low-end PC is plausibie around $200 ASP within the next five years (with a tablet device on average at $260). We bolieve the ablity for vandors to deliver credible PCs at low price points will dive strong elasticity of demand and penetration 3) Consumer PC growth of 17% (including tablets) long term. We forecast PC volume Growth of 18%) 18% in 2011/12 (3% 2% ax-tablets) and 17% long-term (2% ex-tablats), tiven our view that PCs can be delivered in volume at price points as low as $200. This ‘Wo think will drive the installed base for consumer PCs to 1.2bn LT from 680mn last year. 4) Commercial PC volumes to show robust growth in 2011/12, driven by @ corporate rofrosh. We estimate that the average age of the installea base is now nearly sx years, which to us suggests that replacement volume wil recover in the near term. Furthermore, ‘our proprietary CS CIO survey suggesis a furtner boost given the transition to Windows 7 and new hardware releases (more powerful specs and chip releases). As PC penetration fof the labor force continues ta increase, we forecast commercial PC volume growth 17%116% in 2011/12 (13%10% ex-tablets) and 11% LT (5% ex tablets). rardware nCarprrsuss™ “16 March 2011 £5) Overall PC units to grow 18%/17% in 2011/2012 (5%/8% ex-tablets) and 14% long term (9% ex-tables) driven by @ move down market. Long term, we believe that the PC market wil rise in volume terms to 7412mn from 368mn last year driven by a shit down market. In fact, based on our pricing analysis, we estimate that some 67% of volume or 475mn units will sll below a price catagory of $500 by 2015 (versus 31% in 2010). This combined with increasing competition means that we expect ASPs of $675 in 2010 to decline to $415 long term (CAGR of .9% versus -7% historically). However, despite this pricing pressure, wo expect long-term PC revanue growth of 4%. 6) A shift towards emerging markets. We expect emerging markets to represent $5% of PC demand longer term (versus 46% in 2010), driven by increasing penetration per capita, and improving affordabilly as devices continue to become less expensive. 7) Tablets are dtforont from netbooks. We believe there are several reasons why tablets will have @ more meaningful impact on the PC industry than netbooks historically. These Include (1) optimization for consuming media, (2) functionality that can be scaled up (3) a mobile OS optimized for the smaller form factor, (4) leverage from new cartier channels and (5) instant on! longer standby. So far we have observed a high level of usage similar ‘0 smartphones, with significant interest in each of consumer and corporate environments, 8) The PC value chain and industry structure may change. We estimate that nearly three quarters of the PC industry (supply chain) profits accrue to the Wintel duopoly versus a mere 20% for the top five PC vendors. in contrast, for the handset market (where softwarelchipset are delocalized) the top five OEMs account for some 90% of industry profits. We expect the tablet market to emerge with a profit structure in between these two markets, and for this reason believe PC vendors will have a once in a generation ‘opportunity to break the Wintel duopoly and capture a higher value. 9) Tablets t0 represent a $120bn market fong term. The tablet markat is inherently challenging to forecast given its recent introduction. Fer this reason, we adopt a price-point based approach which assumes that the demand for a given level of computing necessity can be approximated by price level. We use our global PC forecast by price point to next dotermine the addressable market ‘or tablets based on a penetration analysis at each of these tiers. We for instance, assume that a low-end PC at $300-499 can be better served by tablets (49% of LT PC demand) versus the high-end $1,000-plus (we assume 2%). We Conclude that the tablet markot could represent a $120bn market by 2015 with units reaching 298mn (or 42% of total PCs), Figure 67: We Expect Tablots to Account for Nos ly Half ofall PC Shipments Less Than §499 by 2015 PCs % by price band “Tablet % by price band “Tablets by price (mn) 201020146 _2012E 2015 _2010_20NE_2012E 2016 _2010_2011E 20126 _2015E Sess ok 8% Ik 19% Oh WH Tk ame OS s300-499 ce ee ee ee es 3500-8699 33% 40% 28K ITH HOH 2am $700 $999, 26% 21% THOU TO HH 1.000 sok 1% 94 kw Totar 00% 100% Too 100% SK 18% zou 42% «7 ~~ Source: Company date, Croat Suisse estates 10) The traditional PC industry wil see revenue declines. Givan our assumption that tablets will represent 42% of PC demand longer term, this inherently drives a significant level of unt and revenue growth, However, excluding tablets, we believe the traditional PC ‘market wil only grow 3% in unit terms and actually decline 6% in revenue terms, rardware wncreprrsusst™ ‘sows ‘9d 1970 Jo zy Bunvosoxdar—17 wwEEe 02 ur wup, | jauigg yo suowdiYS ix ayy *yunpoddo jajqe} 961~7 ‘ejoKo seak-<) oz pieMo} pUaR rian wovBuo} a 1 a ‘0h oF Hau 914 190 SA OMT sever svoneodion pu seu unsioo jan ul Aagepioye Buseaiou sa) aujoap oj enuguco Sas Od SY 24 dn seI9poul & MOYS O} ape1 JUoWaDE| day (E ‘quewaro1duy souewoyed panuquos In jo %SP powUese;ds! ply) si ur yytos8 GuoKs enup 0} yeu ‘Aq vanup yywoIB yoYseW JoWNsUOD (| hq UBKU (%L-+M4914) zunboz ul swum uuzos juweey 2q_ MM (je}a10WWOD pue soWNsUCD) yoHeW od TEIN OY BEWRSD OM, Pa SoIRTO SRT TOTRTTSST ZLILLOZ UI %LL+N%BL+ SHUN Od BALA ©} SHOALeW [eIoJOUWWWOD pUE JOWINSUOD O49 U YIMOID YSNGOY Toad OMA '99 21ND ™Carprrsuss™ Consumer vs. Commercial, Desktop versus Mobile ‘Our approach to modeling the PC industry is based upon the fundamental belief thatthe driver for each marke, ie., the consumer and corporate markets, are quite diverse, and for this reason, we choose to model these segments of the market separately. For the ‘consumer market, we employ an affordabiliy-based approach, whereas in the commercial segment our model is driven by penetration of the labor force and corporate refresh cycles. Within each of these markets we observe a clear industry shift toward mobile, and as such, forecast these segments as two distinct pieces. PCs: A Debate Around Whether to Include Tablets Before we begin discussing the fundamental drivers to our PC forecasts, we believe itis necessary to define what constitutes a PC. We refer to Gartner's defnton below: “APCs @ goneratpupase comouter tats astinguished tom other computers by its adherence to hardware and sofware compattily, APC system i vowed asa single unt which includes a CPU, rman, 4 mouse and keyboard. Furthermore, Garner does nt include tin ciont torial nthe PC classification. Ti eteyory ncudes desk based PCs, alivene PCs, miaHPCs, oer desk-baced PC tors suchas whto-bor and solFassombled PCs, mobile PCs suchas notbooks and tablet PCs.” Interestingly, while many reputable industry sources have included tablet PCs in their definiton of a PC, they excluded media tablets, For example, Intel provides PC unit forecasts excluding the media tablet markel, Furthermore, Gartner has recently cited media tablets as being a key reason for PC weakness in 2010: ‘Worldwide PC shipments grew ato lower ate than our projection [.} Over, holiday PC sales were weak in many key regions duo f the intonstying compettion in consumer sparing. Media tales ‘and other consumer electronics devices, suchas game consoles, ll competed against PCs” Here Gariner defines a tablet PC as a note-book style device, presumably based on the x88 architecture and a media tablet as follows: Modio teblots are “computor processors In @ tnbiotsize form factor, which are optimized for cammuneatons and media consumption, Theie processors ore more gowortul han previous {generation ables, but ees capable than the sama/equlvalent generation of apps or PCs." \While debate sround content consumption and creation on media tablets will continue to evolve, we believe that in large instances media tablets (which heretofore will just be referred to 8s tablets, with tablet PCs being included in our notebook PC estimate) will st the least, become a substitute at the low-end of the PC market. Indeed, even within our ClO survey we found that there remains significant corporate appetite for PCs. We think this defintion is an important distinction to make, because once tablets are included in our forecasts a very diferent perspective of the PC market evolves: 6Carprrsuss™ Figure 89: YIY PG Growth Higher This Next Year When Tablets Are Included “| ‘Source: Company dole, red Suisse ontnales, Garner The market is not slowing. As shown in Figur 88, while we expect the PC market ‘excluding tablets to experience @ deceleration in positive unit growth through 2015 (at & 3% CAGR versus 11% from 2005-2010), over the past few quarters, once tablets aro Included, we think that market growth has actually been quite resilient. Infact, as shown in gure 89, while traditional PC volumes in the last four quarters are stil up 13.8% yoy, when we inchide tablets this increases to 19.4% yoy growth. The difference only becomes. further pronounced going forward First T 3t Launches in 2010, Smartphone YIY Has Accel i I : } ‘Sears: Company data, Creal Slese ealinales, Gartner iN the tablet market cannibalize the smartphone market? As we demonstrate above in Figure 90 smartphone unit growth has begun accelerating despite the launch of tablets like 16Carprrsuss™ [Apple's iPad (April 2010) and subsequently Samsung's Galaxy Tab in 410. This we think Is a strong argument against the view that tablets have begun and will continue to ccannibalize smartphone unit sales. Furthermore, from a form factor perspective we think smartphones as ultraportable wll inherently occupy a separate use case! environmen An Econometric Analysis for the Consumer Market indamentally, we believe the core of any predictive consumer PC market size analysis should be based on strong linkages between key economic variable like price, GDP per hhead (income proxy) and penetration per capita. Following this analysis, we then think it ‘makes sense to have the debate around which PC form factor will garner the most momentum of taka matket share. With this a= our staring point, we ran several fegressions among these variables with the dependent variable in each cage being penetration per capita and conclude that the relationship is statistically signfiant. We would highlight three main points: R-Squared Values of 70%-Plus between PC Penetration per User and Affordability. Given cross sectional data for 42 countries over 8 years (see Figure 91), we find @ statistically signifcant relationship between PC penetration per user and ASP/GDP per head (which we use as a proxy for affordability). in fact, R-squared values for the four rogressions we use in our econometric model (to account for geographical bias) range from 71%, used in our developing market regression analysis to 87% in the US market (Gee Figure 92.) Figuro 81: The 42 Countries Used in Our Regression Analysis Represent 70% of the Global Population and 88% of Global GDP Country Pop im) GDP (son) Pop na) _ GP (oa) {Tagentne Be 17 eH 2 Austral 2 1024 25. Now Zealand 4 va 3 Austia a 39526 Norway 5 381 5 Bal 193169328 Philippines a 413 8 canat 4 1st? 29. Porugal 1" 238 1 Chie w 5730. Russia 40 t28 8 chine tot 5508 at Singapore 5 210 9 Colombia 242-2 South Aca 0 296 10 Denmare 6 318 33. Spain 46 1.469 13 Germany 2 a4s0 35 Taman 2 ae 14 Groce " 318 aT Thsilens oo 208 18 nia 1216 1367 39 UntedKingcom 62 2216 17 Indonesia 235 572 &_Unied States a 14492 18 trlene 4 2221 Venezuola 29 321 pan wr S212 21 Korea 6 289 Tos 22 Malays 2 208 toba set 23_Mexeo 109 oe ef global 70% 7creprrsusst™ ued yetnuog fe pura 200019 ‘urbjg (eo) 8) papnjul SuyUNOD ‘wepbuy poHUN‘saEIGPoHUN ‘WEN ‘pUELEZNS “uapang ‘erodeBus ‘Aeavon pues? man "spUEVOUIN B9!0% UECEr ‘BJO BuoH“Aveiog SUEY BBA nLed "BOBUED “sy "BEENY {IO} GL] PANU SaLAUNED sojeuyse assis aP819 e1ep fuedued ‘ames {6 sued ep 0 GUNN zk uodeUop IS ToRSSaIbST SaaS POHUT “JOHLEUI dd 2MN 10} japous pueUIEp WHHa}-Buo] e AUP (01 pasn oq ues Yatym SUN pue awooUI/dad ‘dS¥ Ueerneq Hull 4v9j9 © s1 o10y} aAaiioq “gueroyjaoo uoisseiBe1 fp 40} eouroyuGis jo jana] yBy e y0B6ne GI< Je SONIA Ise, ~ 9416.01 % LL, wow Bubues 5,4 yum ‘erep eM ysBuoUE Luogej21209 Jo jane} YBIY ® enoid suo|ssei6e1 nO “sone a UBIN ~ im 400g prema ase oe “ (eyde9 10d ya froud v se 10 ewoou) Ayyqepioye pue eydeo sod uogegoued Dy voonjoq diysuoneja: [UEWepUTY eM UO Paseq jepOW Oq POseG-OLNeWOUODS JS) OY '9q 0} onayeq on yoy pedojanop aney 9fA"|POW DUFEWOUOD® Sil ~ ‘ABojopoujoy PUE [2PON Dd OU}eWOUOD; INO Uo pesf, suo|sseLBey Jo AreWWUNg :Z6 21615 ‘sisfleue uo|ssai601 no uo poseg < ™ ‘907 sud wep Jo equ ‘Were uoNeNap PIs exons 18Carprrsuss™ The Elasticity of Demand Is Greater than One Using quarterly PC unit shipment data and the corresponding global average ASPs, we have platted a simple demand curve for the consumer PC market. (See Figure 93.) This, then allows us fo estimate the incremental volume impact from a unit change in price. In fact, while demand elasticity for global consumer PCs stil remains at 1.2 (as of 2009), tis, has come down over the last six years from 3.1 in 2003. That said, the global elasticity remains above one, which suggests that any Industry price cuts wil generate relatively higher volumes and actualy be accretive to industry uns Figueo 83: Demand Elasticity for Global PCs ts Greater than One $1,300 $1,200 $4,100 $1,000 208 Baatioy 20 ASP (USS) 900 $800 $700 600 ° 5000 10,000 15,000 20,000 25,000 30,000 Global quantity demanded quarterly (000) Price elasticity of demand for PCs on a global level is stil greater than 1, which means that a further eduction in ASPs will load to a mor increase (additional 20%) in quantity than offsetting 40,000 Source: Company data, Crest Suisse estimates, Garner ‘So How Do Wo Use This to Be Predictive? (Once the core relationship between GOP per capita, pricing, and penstration is testablished based upon the regressions shown in Figure 92 (which we again highlight to make the distinction between developed and emerging markets), forecasting becomes ‘more straight forward on a long-term basis using the core inputs of ASP decline by country ‘as well GOP and population growth. Given GOP and population growth rates over the long term prove to be rolatively stabla, the key assumption driving our forecast, therefore tends up being how the ASP of a PC may evolve, How Low Could PC Price Points Go? $200 [As discussed, we have found that PC penetration for consumers is closely linked to the ASP of the device. While the dectine in component prices, and hence the BOM of the device is clearly the largest criver of ASP declines, equally we've found that new device categories, like netbooks, are further able to penetrate lower price points given lower 19carorrsuss™ tomva sor specifications. Taking into account these data, one thing appears to be consistent: ASPs for the consumer market have seen a relatively stable decline of ~7% over the past five years, Furthermore, this trend has been observed across all price points as demonstrated in Figure 94, While in 2004, only 3% of total PC volume sold for less than @ $500 ASP, this, number last year was significantly higher at 31%. For this roason, we can conclude that the long-term installed base for PCs is ultimately determined by the retail ASP of the PC at least atthe lower-end of the market. Figure 96: A Shift Toward Lower Price Points Has Been a lear Trend Over the Past Eight Years 100% 4 ‘See aati vars 2004104 90% +1) the cumuinive pacontage af nia pool vse Under $00 ASP anevada 21% a 2 2) An cutive poreniage oft gba ‘ume under # $1000 ASP nas meranres 087% 70% 60% 50% 40% 20% (Cumulative unit share under price point, % 20% 10% o% Handset market We observe a he handset mae tok on verge ay a yore (ersus ou suman fe fe Be PC marl > smariphone mark. Wve cbserve at nb sraipane metel whch ony mead in the ery pa fhe a! decae, al penetation per ‘capenineased fom ~10% 0 ~22% in around eo years ‘Source: Garner, Company date, Groat Suisse tates rardware 88creprrsusst™ s2jeuyse oesins apei9 ‘Hep fusduog ‘ames s1ok 550 2peg SuIP/O a10}eq (2102 a steak ¢ Aq way 200 ‘249 auIpep 0} Soy fers1ouINOS 19) op qwouee\do1 ay ack an "2 SMODUM 40 Uogdope a.) puB souyoew eouEWOLed souGy fq ven Buea 81 uOKI ysayes osodioo iueuno og cy SUMO ‘yondn ysapou v 298 0} yUDUIsOEId—R (E 4510 sion jeouors (suun yo gg) 17 syun umzy pue zy s402 Ur sun uUljzalug yosaidar 0} 0196) oodxo omy 'Sie}9e} 204 Ajunpoddo y (Z az'L 0 $94 [es8UNo9 20} 980q pajeysu) 17 © 1990x994 “2010) 1oge} |ebiewuco 214 jo SUR uoKeLeUed ‘24 us sasearou! ysopou uanB (AreouOISHY ‘syun od yodxe am 17 “(uazersmaezis i) Poued si) 10n0 YesAPe Jo \yno16 onuonas yim (uueycruMe0E yo Syun) zLoz UL ole MoLke 40 «ymou6 yun O¢ [eo1eWUIED JooKxo ont Pa SoS ST OITTSST (sun wuspos O17 %b4 Pue (siun uurzyZ /uuiBOZ OH ZEILLOZ Ul NOL +I%LL+ MOID OF SD¢ [eIDIOWIWIOD :Z0} O:Nbis 29Carprrsuss™ “16 March 2011 Corporate Refresh Is Two Years of Robust Growth lowing our discussion of the consumer PC market, we turn our attention to commercial PCs which represented 48% of global units and 50% of global PC revenues last year. Unlike the consumer PC markel, in the commercial market, sales are more driven by responses of enterprises to new hardware releases (more powerful specications, new chip releases), software refresh as well as by the steady penetration ofthe laberforce as productivity increases. For 2011/12, we forecast units of 208mr/242mn (+17%/+18% yoy) and 11% longer term to 301mn driven by several factors: Installed base to see steady growth of 11% LT. According to Gartner estimates, the installed base of commercial PCs is currently around 700mn, and has grown ~10% over the past five years. Going forward, given (1) continued productivly improvements in PCs (higher specs at lower prices), (2) the industralzation of emerging markets, and () consistent and increasing penetration gains of the laborforce we expect the installed base to grow 11% per annum through 2015 to 1.2n users LT. Figure 103: Long-Term Installed Base Growth of 11% Modestly Above Historical Averages. 2005 20067007 700870097010 20TTE 20T2E 20TGE 20T4E Z0TSE _CAGR 200510 CAGR 201056 Asia Pcie so 152. 172 195 219 247 280 918 6244 47S TRON 10% Wostem Europe 9 108 14 119 129 tanta 169179 79% 60% Easter Europe a7 42 49 58608878 108 7 130% 15% MigdeEat&Arcs 22 2831989 a 68 3% 95% Latin Amare oo we 7 8 @ s09 14 187% 14.9% iota 57a s31 503630708 7a} 80595758 Ua7T WoT ToT “Sowa: Company data, red Sues eutimates ‘Adoption of Windows 7: stil inthe early innings. While Windows 7 has only launched in October 2009, itis already one of the fastest selling versions of the Windows operating system. n fact. at the Consumer Electronics Show in January 2011, Microsoft CEO Steve Ballmer noted that 20% of PCs now connected to the Internet are running Windows 7. Given the ‘gap’ seen in many commercial enviconments, whereby corporations remained with Windows XP despite the release of the newer Windows Vista, we expect an upgrade ‘0 Windows 7 which is being met with improved functionality and performance from hardware vendors. As shown in Figure 104 and Figure 105, according te the Credit Suisse IT Survay, the Windows 7 upgrade is only now beginning to take place at enterprises, with only 27% (on average) of enterprises meeting the target completion date between Q3/4 2oit rardware 90Carprrsuss™ “16 March 2011 Figure 104: Enterprises Are Still Only 27% Completed on Figure 105: Enterprises on Average Expect to Complete ‘Average with Their Windows 7 Upgrade the Windows 7 Upgrade between Q3/Q4 2011 ‘Question: How complet is your Windows 7 intograton? ‘Question: When do you expect te complete your Windows 7 Integration tasting arpa date)? ‘An aging installed base of PCs, We estimate that over the past fow years (see Figure 108) the average age of the commercial PC installed base had been between 5-6 years, some 1-2 years longer than historical ages owing to a slower macroeconomic environment and therefore reduced corporate spending Figure 106: We Expect Average Ago of PCs to Decline in 2011/12 Owing to the PC Refresh, Then Slowly Rising LT 200620072008 20082010 _2011E 2012E 2013E 2014 2015E Avg 2006-10 Avg 2017-185, Asia Pcie 59 57 62 67 63 60 48 60 52 53 67 51 Wester Evrops 40°39 42 48 86 50 47 48 49 49 45 49 MigdeEataArcs 59 5.1 54 58 62 S57 SS Se Se 60 5s 87 Latin Ameren 545261 66 107 89 62 66 90 98 68 89 Giowal 4545 808358 53 50 52 5458 50 53 “Scure: Company dala, reat Sues extmates. \We would highlight that at this age, most PCs begin to fal and the actual cost of repairing such devices on the client side can be prohibitively high. For this reason, and as corroborated by our Credit Suisse IT Survey, most enterprises are planning a ‘major’ PC refresh for their businesses (on average) at the end of 2011 as demonstrated in Figu’s 107. rardware soCarprrsuss™ “16 March 2011 Figure 107: Enterprise on Average Expect to Upgrade PCs Around the End of 2011 ‘Question: When do you expecta maj PC caraen? ‘Source: Great Suseo IT Sur, Developing market units to more than double over the next five years. While we expect the Windows 7 refresh to postvely influence demand both in developed and developing regions, we think @ more significant structural trend is the continued use of PCs in ‘emerging regions, In fact, we believe that over the next five years, PC units in developing regions wil grow at a 15% CAGR, versus a CAGR of only 6% forthe developed regions, Figure 108: Commercial Unit Volume in Emerging Markets Will Exceed that of Developed Markets in 2010 and LT USS n ions, unless otherwise sated Devetpea markets or 3 3 710 12 ee 0% Deveroping markets 20 92 m 32 ee 15.9% Tovar 6 178208 2a as area 0% Source: Company deta, Ort Suisse estates Performance remains a key consideration. One simple driver of the PC upgrade cycle is the evolution of the system and processor performance as affirmed below by an IDC ‘survey of commercial PC buyers conducted in 2008 (price being the second}. rardware 92Carprrsuss™ Figure 109: IDG's Commercial PC Buyers Survey Reveals that Performance (61%) and Gost (3194) Are the Two Core Drivers ofE Desktop virtualization, @ potential negative headwind. The idea of a server-hosted desktop, which effectively consolidates the computational power of individual desktops. and instead replaces the user's machine with a thin cent, could prove to be a headwing for corporate PCs in the intermediate term. If desktop virualzation were to become @ rend. in enterprises, whether for securiy or cost reasons, this would in theory put downward pressure on the sale of traditional PC units (and ASPs given decreasing need for performance). However, this threat may not be imminent in the near term as suggested by our Credit Suisse IT Survey (Figure 110), which indicates only @ modest rise in expected desktop Viualization (lo ~25% of machines) two years from now versus today (14% today) 93Carprrsuss™ “16 March 2011 Figure 110: The CS IT Survey Suggests 25% of Desktops May Be Virtu 1, in oth ot ized in 2 Years ‘Question: Wn nt of your dest ont in ne? ‘Source: Great Suseo IT Sur, Tablets a $120bn Question Long Term While the PC tablet concept has existed for nearly ten years (think laptop with swivel screen), the “tablet” market as we know it today, which involves a fingerinput on touch screen (which i the way we define the category), was only brought mainstream by Apple's IPad in 2010. This introduction by Apple sparked the interest of mobile industry peers as Well as traditional PC hardware vendors (the tablet in fact is a convergence between the ‘mobile and PC industries), In Figuco 117, we list afew of these key vendors, Figure 11: Sample of Larger PCiHandsot Vendors Entering the Tablet Market Vendor Opera we rating x Tani Windows we ‘web0S! Windows ove ios Lonove: Ando! Windows nrchos ‘Arvid Windows is ros ‘vaya ‘kis Panasonic Androis et ‘oid Samsung ‘droit Windows ken ‘Avs Viewsonic Anois Source Company del, Crest Suisse estimates ‘A further complication is whether tablets, are cannibalistic or additive to the PC industry While intially many industry participants expected media tablets to be niche products, we would argue thatthe usage case continues to evolve daly. This leads us to believe tablats ‘may meaningfully affect the PC industry and more so than netbooks, which appear to have already peaked in 0409 at around 10mn units, (See Figure 112.) rardware 94Carprrsuss™ Figure 112: Global Netbook Sales Appear to Have Peaked in Q409 Unis miions,unibss athens sated ‘Scwee: Company data, Creat Suisse estimates. ‘A few reasons we think tablets are fundamentally diferent devices than netbooks are the following "= Optimized for consuming media. The tablet form factor allows the user to hold onto the device at a comfortable angle for viewing video clips, movies, playing games and reading. This compares with the clamshell form factor of netbooks which inherently lends ise to being set on a flat surface, "= Functionality can be quickly scaled up. With the exception of the iPad, most tablet {devices (and we expect nearly all in the future) will allow users to swap out for a larger memory card, and add @ keyboard if necessary (a8 opposed to having it permanently attached to the device) = Mobile OS optimized for smaller form factor. We think Windows Is optimized for larger ‘screens, but when it comes to sizes 10" and below (which is addressed by netbooks and tablets), we think the user experience should change to accommodate the reduced scraen real estate. Tablet OSs liks iOS, Android and webOS were created to ‘optimize the user experience at smaller screen sizas, versus Windows, |= Instant on and longer standby. When it comes to checking mail, browsing a website or taking notes/pictures, users inherently prefer this functionally to be responsve—which is made possible via Nash memory. A mobile optimized processor (ARM-based so far) and mobile OS allow for standby of up to 30 days (as see on the iPad) versus at most ‘a fow days on a netbook, ‘Corporate uptake and intorest appears meaningful. While we believe products like the iPad have been initially designed with the consumer in mind we would highlight that the level of corporate interest has been surprisingly nigh. Infact some 80% of the Fortune 100, are deploying or piloting the iPad according to Apple's CFO, Peter Oppenheimer. “Enterprise Ci0s are adcng iPad to thar approved dove Ist at an amazing rato. Today over 80% of the Fortune 100 a6 ateady deploying or pilting iPad, up Hom 85% i he Saplember quater. Some recent examples neue JPMorgan Chase, Cardinal Heath, Walls Fargo, Achar Danae Miland, ars Holing ard DuPont” — Potsr Oppenheimer, Apple, CFO In Figure 173, we note a few instances of tablet trials in large corporations and a few school districts. We would highlight the diversity by which these devices are being used. 9Carprrsuss™ “16 March 2011 Figure 113: Various Announcements Indicate Strong Interest for Tablets in Enterprise and Edueation Malte Foanc Payoe San fe Fanci Pypsoe ital oie cf 100 rpc SAP sears Pybooe Tel 3205 Co Supety Chan Denver nematond Argo 100 puyeese Seca Opeaton ley p36 ee Corde Dearne of Carectont Payson Lake rave nseperdent Sel Dit Payson ‘Soure: Company dota, Credit Suisse estates (Our Credit Suisse IT Survey equaly ilustrates strong interest in corporate tablet adoption 15 represented in Figure 114 and Figure 115, and shows to 30% of PCs LT could be tablets rardware 96Carprrsuss™ Figure 114: Mobile PCs Like Notebooks and Netbooks ‘Stand the Highest Chance of Being Replaced by Tablets ‘Question: What ithe probably that a table wil replace the flowing Figure 115: Which Our Survey Suggests May Be Upward of 30% of Total Commercial PC Demand ‘Question Wnt ofyour global amployee base have/ wi have @ vis inthe next tae years? tablet device (su tod by entre) a those points inte? ‘Apps support from the start. Another aspect of the tablet market which makes it attractive relatve to traditional Windows-based PCs ‘s the number of apps available on the various platforms (generally for a nominal price), as shown in Figura 117. In addition to sheer rhumber of apps, the variety of offerings from books to games, entertainment, education and produetiviy is extensive (and we would argue relatively complete) Figure 116: Apple App Store Growth Has Been Strong _Figure 117: IPad-Specific Apps Have Seen Significant Since Launch, Offering Over 325.000 Apps. Growth, Reaching Over 65,000 Apps by February 2011 ‘Source: Company data, Gro Suisse estimates “Source: Gomany dota, Gre Suisse oxtimatos Carrior support opens a new channel. Given key drivers of tablet demand include the noes for instant on, connectivity (WHF but also cellular) and mobilty, the carrier channel becomes @ natural and globally well established outlet for sale to consumers. Although netbooks did try their hand at carier distribution (embedded with cellular connectivity) i the U.S, and Western Europe, and they failed quickly; we believe these devices weren't ‘competing from a user standpoint. In fact, we'd highlight a few core aifferences between these devices, suggesting a more significant uptake for tablets given: (1) always on— tablets are and netbooks are not, (2) mobiity—the tablet form factor nearly twice as thin (G) battery ife—the (Pad gets ~19 hours web surfing and watching videos vs. a typical netbook at merely haif this time, (4) apps ecosystem and cloud-based synchronization for tablet OSs versus Windows, and (6) wider array of tablet choices for consumers, from OS te hardware, rardware 97carorrsuss™ Significant vendor support. Tablets in our view represent the ultimate convergence product, falling somewhere in between a smartphone and a traditional PC. By ‘consequence this has attracted the interest of myriad companies as highlighted above in Figure ‘1, which can develop 9 wide range of products across numerous price points, High usage product. As with mobile devices, the tablet is often being used to replace one's time spent on a computer (which nowadays largely includes web-related activities like e-mail and web browsing). In fact, according to a Cooper Murshy Copywriter survey of over 1,000 U.K. iPad owners, on average an individual spent nearly 10 hours a week (or about an nour and a half per day) on their tablet, which compares with between 2 and 5 hours for smartphone usage based on various surveys published on the internet. Working Out the Tablet Market; Look by Price Point The tablet market has certainly seen fast growth over the last year, but we now address the more important question of how significant the tablet market opportunity is long-term. While there are a number of ways one might forecast the tablet markel, we take a price-point based approach given we think ths serves as a strong proxy for use case (in the following section we provide a cannibalizaton-based approach as a way to think about upside to our estimate). Based on this approach, we expect users to spend some $120bn fn tablets longer term Figure 118: We Expect that PC Volumes Continue to Move Down Market Lon Bm i” 3 gm 2 osm ‘Source: Company data, Croat Suisse estimates Step 1. Projecting the PC industry by price point. For both consumer and corporate markets as discussed earlier (and as seen in Figure 94), we have observed a clear shit down market for PC volume. In particular, we would note that PCs selling under an ASP of {$500 increased to 31% of total volume (a8 of 2010) from a mare 3% in 2004. We then argued that we expect this price trend to continue through 2015 as (1) new entrants (handset vendors primarily) enter the market, (2) BOM continues to decline, and (G) vendors move to address the emerging market opportunity. As such we expect that by 2015 (shown in Figure 118) cumulative PC volume under an ASP of $500 increases to nearly two-thirds of global units (67%) 98carorrsuss™ Step 2. Projecting out the desktop and mobile markets... a theoratical exercise. Before we move on t0 the percentage of the market we think tablets will occupy, our approach requires that we estimate how large the desktop and mobile PC markets could Be, by price point, had tablets never been introduced. We acknowledge this is somewhat of 8 {theoretical exercise, but recall thatthe fundamental basis for our model is agnostic to form factor. For this reason, we can estimate the price points which tablets are most likely to ‘occupy based on use case (discussed in step 3). Step 3. Tablets will cannibalize only certain price ters, but stil see volumes of 298mn longer torm. Now that we have an estimate of the size ofthe overall PC market by price point we believe a framework for thinking about the potential size of the tablet market becomes more straightforward. As shown in Figu’e 119, we believe that the markat for lower-end PC market ie., in the price range of $0 to $499 will see significant penetration by tablets long-term in the range of 45-50%. Tho contrasts with less penetration of overall, PC units in higner price bands, lke $1,000-plus, where we think tablets only represent 2% Cf global volume. The end result based on this approach is that tablets can represent a {65ma/ 1416mn unit market in 2011112 and 298mn unit market in the long term, 99carprrsuiss™ semaren 208 Figure 419: Tablets to Represent a 298mn Unit Opportunity (42% of Total PC Units) Long Term ‘nitions, unless othorze sates We forocast tablet units of 6mmn/116mn In 2011/12 and 296mn LT representing 2% of overall PC demand, Here we have used a Drce-point based penetration approach to forecasting the market glen a strong argument for use case by pice tle. 1) Higher tablet penetration at lower pric ters. We think tbls offer the ues alse expencive aerate to web browsing! eral ard ‘ter connect than a deehop ormobie PC. While we expec thisto change overtime as tablt OSs become moe feature rh an hargwareepecs nprove, we expect igh penton in low PC pices, Between $800-§499, where we expect he highest rumber of {abt devios tobe sod, we assume 24% 38% perevaton m 2011/12 and 40% ong erm. 2) Consumer versus commercial. White tablet device crsiny does and wil hae egnifcaterterpriteeppicatons, we continue to belave abats ae best suited for consumers who place a hier valu on always on ua portable i connected devices. We assume thal longer-term abate represent mere than al of consumer PC velumes and 26% of commercial vekmes (42% combine) 13) Mobile versus state. Given tablets adress tho mobile sognantcftho PC market more han the static desltop maka, we assume higher penevation of moble PCs (54% of glabal lunes) versus ony 11% for destops, ‘Source: Company daa, Groat Suisse extra. rardware 100s2jeuyse oesins apei9 ‘Hep fusduog ‘ames ‘opisdn %@4 s0uany © sondua! syn SV 1919 oBei0Ne 17 947 UO posea) ‘syn uwgg feuogippe ue oy BuIpoo “WEY JO UOHEZIIEQUUED D]UO:;919 Jounsues yeione 5806805 stUL siafelg ana amrowoany 1 %056 sofa ana 1aen6dJ0 1001+ 308409 Buluieg ose soja Bueg ayqeud 4000) « sand 10526 “S]@yrew SoqUos|DO}9 JeUINSUOD J91H Bumjequues sen Jo (aprdn) 1edu jad 249 19ps00 o} wELOd i Sa HT RTT ST (1-1 uwiggz snsion) uuigge 10 04614 181-28 Pineg sun JIaeL. 2} rEUINs 17 JNO sys96BNg oye pug 10u10 Jo UOHEZHEGNNUED 921 ound 101Carprrsuss™ Tablet Cannibalization Is Not Just About PCs While our preferred methodology for projecting tablets, as is shown in the section above, involves a combination of top down PC forecasting and price (as a proxy for use case), fequally we think it is important to consider the disruption tablets may cause to other industies. When this cannibalzation of other industies is taken into consideration we think it could add S5mn units to our long-term (2018) tablet estimate of 298mn, implying further upside of 18%. Here we would make the following observations around whether fur estimates could prove conservative or optimistic. Cannibalization of the commercial PC environment. Corporate penetration is stil quite low. [As shown in Figuro 119, we are stil only assuming that within the corporate marke tablet penetration remains low al 26% by 2016, though we would nate that our Credit Suse IT Survey suggests th could be as high as 30% by 2018. Furtner, our proprietary survey suggests that traditional mobile PCs stand the risk of being replaced (notebooks and netbooks). Cannibalizaton of other consumer industries could offer upside. The table is similar to the ‘smartphone in the sense that replaces not only a large proportion of tasks competed on 3 desktop or laptop, but it could also cause aisintermediation of use for other product catagories like the below. (For details please rater to Figure 120.) © Eveaders. Given tablets large screens of 7 and 10°, relative to smartphones, it would ‘seem that @ natural extension ofthe device would be e-eading. In fact, according to @ recent Cooper Murphy Copywriters poll of over 1,000 U.K. iPad owners, respondents preferred reading newspapers, magazines and books on an iPad relative to all ober ‘devices or physical forms of the media. (See Figure 121 and F gure 122.) We estimate that nearly 100% of the e-reader market could be cannialzed by tablet devices by 2015 (from 25% today) Figure 121: The Highest Number of Respondents (31%) _ Figure 122: And When Asked About Preferred Method for rofor to Read Newspapers and Magazines on an iPad Reading Books, Respondents (41%4) Again Chose the iPad ‘Source: Cooper Murahy Copyanlers poo ,094 UK Pad owners ‘Sauron: Cooper Murphy Copywrera palo 1,094 UK Pad owners "= Portable DVD players, TVs, MP3 and other CEs. As highlighted in Figure 120, there ‘are a number of devices that currently fil tha need for on the go entertainment, ranging from MPS players to portable DVD players. While tablets only address 2 {action of these devices given limitations on connectivity (external ports on the tablet ‘wireless. connectionsistandards) and device content, over time we think it makes sense for a user to own one integrated device, For PNDs, gaming consoles and laulometive DVD players we assume LT cannibalization of 25%/35%/50% and ‘complete cannibslization for portable gaming devices and portable DVD players, 102Carprrsuss™ "= High-end smartphones. While our analysis of additional consumer electronics (versus, just computers) in Figure 120 focuses on products with single functionality, given the tease of consoldation into a single platform like a tablet, equally we would be remiss not to highlight the possibilty that tablets cannibalize ultra converged devices like highend smartphones (taking into consideration the evolution of technologies ike \VolP). Given our estimate that over 130mn smartphones wil sellin the high-end by 2015 (>$500 ASP), even 25% cannibalization of this market would suggest a further 10% upside to our current LT tablet forecast Figure 123: High-End Smartphones (2$500) Will Represent Over 130mn Units by 2015 “Smartphone market (mn) 350 ° 0 s 7 6 a 80 $50- $100 0 0 88 tw 100-5150 0 8 mw 3150-8200 ee er 5200-250, cr 250-300 2 5 ww tH 300-350 1 wg 3350-400 non mM om 409-450, ee ee ee ee ee ee ee 450-500 2 ek > $500 Pe “Souee: Company data, Cred Seve estimate, What's next for tablets? The X-factor to forecasting the growth trajectory of this market LT Perhaps the most challenging aspect of looking at how tablets can impact the overall telecom and consumer electronics market is considering all the future applications that can be addressed by such a device. While we nave tried to capture this in the above analysis, inherently we believe our estimates could be conservative as they ignore newer applications. Good examples include (1) using the tablet as a remote for the home, (2) as ‘a menu at a restaurant, or (3) as a point of sale device al convenience store or & pharmacy. Admittedly, these are niche markets for tablets now, but we argue they shouldn't be entirely ignored. Once in a Lifetime Chance to Break a Monopoly \Within the PC industry the Wintel duopoly is hardly a new phenomena, howaver wo fin it is worth exploring in light of new PC form factors (particulary tablets) given hardware vendors’ opportunity to use new suppliers and effectively break the strangle hold Microsoft {and Intel have on the industry. AS shown in F gure 124, Microsoft and intel (Wintel) have accounted for over 75% of industry operating profits last year. We would highlight the following key reasons the vendors representing the other 25% of operating profits in the industry would be interested in supporting a viable alternative (1) ARM-based chips loading for now. ARM-based tablet chips are gaining signifcant action versus the x86-based Atom processor given (1) a power management advantage currertly, which cannot be over stated on a portable device and (2) support irom leading OSs tke Android, (OS and webOS. While Intel expects its ‘Oak Trail tablet chips to close this gap (we would note that during CES in January 2011, the company highlighted design wins with over 100 tablets), product timing and viability (consumer uptake of Windows 7 tablets) remains fo be seen, We'd further note that key ARM-based suppliers like Qualcomm, Apple, Samsung, NVIDIA and Marvell are already designing dual-core processors (some even quad-core) which wil further the narrow the performance gap. 103Carprrsuss™ Figure 124: The Operating Profit Share of the Smartphone Industry Is Much Different than that ofthe PC Industry Given the Nonexistence of a Chip/Software Duopoly (Wintel) ws 1th tr ony _ om & aust Ae sven, Merosatalon Bon oy i _ Savoy eso tone o 30% 10% of ndustry profits, vs. over 40% for ‘Scare! Company data, Grea Sues estate. (2) Microsoft Windows not optimized for touch-screen yet, Based on our conversations with industy experts and former Microsoft employees, we believe Windows 7 simply cannot be optimized for touch (in terms of finger input on capacitive touch screens). Here we would highlight a few key points = The company has historically focused on integration of touch around handwriting recognition! stylus (versus finger touch), = Most Windows applications are written for a mouselfine point (je. stylus) and therefore are not conducive to use with a finger (capactive touch). = Microsoft tracitonally optimized thelr software to function as a platform which would support all hardware-types. Given the tight vertical integration of hardware and software vendors in the smartphone and tablet markets, the company may need to Fedefine its target hardware suppliers. A good example of this is support of only Qualcomm-based Snaparagon chips in the new Windows Phone 7 smartphones. = Windows 8 (which the company noted will also be compatible wih ARM-based processors) will ikely be optimized for touch, but this we think willbe released in 2012, Given the current x86 platform is not optimized for battery life, this makes for vendors Using Windows on a tablet even less convincing. Here we'd note that the Pads ARW:-based application processor has a battery lfe of ~10 hours on normal usage, but {an Wintel based tablet on average may only ast for 5-7 hours on average. 104Carprrsuss™ (9) Vendors can pick and choose OS and chipset—more industry competition. The ‘wo clear winners from a aisaggregation of the Wintel duopoly are PCihandset vendors and consumers. First PC and handset vendors can choose from a varying number of [ARW-based chip suppliers including Qualcomm, Samsung, NVIDIA and Marvell but also (OSs like Android. As innovation accelerates to fil sockets, consumers sland to benefit fom enhanced technology at competitve prices. (4) Google offering close integration with hardware vendors on key launches. One key difference between the Microsoft and Google approach to ther respective operating systems! hardware partners (Windows and Android, respectively) is a platform versus targeted strategy. Whereas Microsoft has taken the platform approach, whereby any vendor can plug and play with the oparating system (and therefore qualityiperformance is, left up to the partner, in our view, Google has taken a more focused approach by working closely with key vendors to provide an optimized hardware solution. This can be evidenced by new Android launches on key hardware devices as shown in Figure 125; for ‘example, Google worked with HTC and Motorola to provide the highest quality experience for the consumer in the flagship product launches. Figure 125: Google Worked Closely with HTC and Motorola on Key Android Launches rarid 10115176 HTC Dream ie. ‘vid 21 HTC (Googie) Nexus One Anat 20 Motors XOOM ‘Scuce: Company data, Croat Suisse estinatos [Al this means that asthe industry shits toward tablets (which we think will represent 42% of total PC volume 2015), Wintel's share of PC industry operating profs deciines as that of other chip and OS vendors increases, What Does a Tablet Need to Become? We acknowledge that our postive view on the tablet market (298mn units LT or 42% of global PC volumes) inherently assumes a number of hardware and form factor upgrades longer-term. In particular, the tablet today which is primarily used for content consumption will need to be optimized also for content creation. Below we note a few of these changes. Improved processing power. As shown in F gure 128, wa see that on average the spaed of {a netbook is 1.6GHz at a price point of ~$350. On the other hand, tablets have 25% less processing power but on average cost nearly 80% more. We expect that this difference ‘moderates over time as the ‘novelty’ device premium for tablets reduces and ARM-based vendors continue to advance processor speeds. In fact, tablets are already embedding dual core processors ata rate faster than netbooks. Figure 126: Netbooks Still Have a Price/Performance Advantage to T ots, but Tablets Are Closing this Gap ‘Top selling netbooks ‘ASP Processor Speed Gores Tablet launche ‘ASP Processor Speed Cores TTosnbe NBSOS| 320099 Atom ASS 17 GHZ 1 Apple Pad 350000 Appa nt —10GHE + ASUS Eee PC 1001PK «$269.99 AtomNASD 1.7GHz 1 Samaung Galaxy TaD «$750.00 Sameung 1.0GH 1 ASUSEeoPC 1OTSPEM $265.54 AroMNSSO 15GHz 2 Blacsbory Playbook $600.00 T4430 10GHr 2 ‘cor Aspire One AOD255E _§26999 AtomNASE 1.7GHz 1 Motorola XOOM 60000 Toga? 10GHr 2 Tosna NBSOS $57999 AomNSSD 15GH? 2 HP Sito 500 $800.00 Atom 540 19GH2 + ASUS Eee PC Seashell 1215N $481.54 Atom 0525 18GH2 2 Dall SueakT* $0000 Tepa2 10GH2 2 HP Min s10-9190NR $90999 AtomNASE 17GHr 1 Cisco Cae 90000 Atom 1 6GHr Gateway L131790 $38900 AMDLs10 12GH2 2 LEG-Sile 360000 Toga2 10GH? 2 ASUSEsePCTIOIMT “$459.55 AtwmNASO 1.7GHz 1 ASUSEae Translomar $490.00 Tagra2 1.0GHz 2 Toshioa Mii NB255 1$22899 AomN&85 17GH=__1_ ASUSEee Ped Nemo $600.00 Snapdragon 12GH2 2 eorage 33586 Teche 14 Average 3s35.00 T2Gke 16 Souce: Amazon com, Cred Suisse estimates. rardware 105Carprrsuss™ Accessories to assist in content creation. While tablets’ functionality nowadays ‘s clearly facused around the touch interface and limited need for outside peripheral, we think this Js more a function of what the device is being used for content consumption), versus what we think it will be used for longer-tem (content consumption and creation). The accessories that generally soll with an iPad today (see Figure 127) are used to protect and connect the device (to a computerimonitorfpower outlet) versus altering the functionality of the device, The exception to ths is the extemalidockable keyboard which allows users to faster input text, whether ibe for email ot notes. This is the type of peripheral we expect te drive increased content ereation on the deviee therefore moving ito become the users primary PC device, Figure 127: Majority of Accessories that Sell with an iPad Do Not Influence the Way the Device Is Being Used Koy ad accessories Price Description ‘Rope Pac keyboare dock ‘368.00 Gonbines a charging dock wi ule eyboars Appa Pad case $38.00 Protects Pad and can be used in various postons ‘Apps Pad cock $33.98 Docks and cnarges Pad, aso cuss 2 aus ne ou ‘ope Pad camera connection $29.00 impor polo and viseos fom gl camara ‘Appa Pad 1091 USB power adapter $29.00 ChargesPac drecty through an electra out! (ong) ‘Anpe Paddock connector o VGA adapter $29.00 Lats you connect your Pad toa TV, moncor or prjector Macaly vowing sand $1801 Improves viowing angle ping (horzera vaca!) Bekin seen protector 511.09 Provides a clear protctve shel, preventing scratches Macaty privacy sereen $2884 Darkane te screen (angles) to prevent viewing by alters (Pag Smart Cover $39.00 Magnetic cover for scean, nkeduced by Ape councton wth Pea 2 eoroge 32.58 ‘Source: Apple.com AA true power efficient mutttasking solution. While most mobile OSs these days have Some form of mult-tasking (JOS, Android, Blackberry OS/ QNX, Windows Phone 7, \webOS etc), the solution is generally (not alvays) inthe form of suspending and resuming he application to preserve battery Ife. We think that longer-term in order for tablets to be a true content creation device, a true mull-tasking solution—which means a user can activaly run a program in the background, whether i's crunching data or a streaming download without pausing another application, and that is also power efficient will be necessary. We don't think the technical hurdle to creating this selution is high so would ‘expect ths limitation to be covered over the intermediate term. ‘A synchronization device to a primary device. All successfully selling tablets (Le., IOS based iPad and Android-based currently), require the user to synchronize the device to @ PC, which can assist the user in adding content, services, updates etc. While this is clearly diminishing over time as over the air updates become more popular and downloading content over 3G (or WiFi) is preferred, equally one less often links up their tablet or smartphone to a PC to offload data, We think that more standard functionality i.e, regular USB ports, may be necessary to transfer data from the device to the PC as the tablet becomes used more for content creation, While currently the Windows 7 based HP Slate '500 is the only tablet which provides this functionality, we expect other vendors add this, feature to their products over ie. 106Carprrsuss™ PC—Winners and Losers \What makes a successful PC vendor? This Is not an easy question to answer especialy given constant change in industry dynamics, including new product categories (most recently tablets), mode of listibution (toward indirect channels from direct) and ‘geographical exposure (strong growth in emerging markets compared to developed markets), ele. Last, itis important to note that ultimately even HP, the market leader in PC Industy (market share at 18% last year), only enjoys a midsingle digit operating margin, \which one might argue is hardly attractive. For this reason, we think execution on multiple fronts remains paramount to capluring value share. In this section, we attempt to score leach of the vendors on six key metrics in order to determine who will win and ose LT. ‘Scorecard Categories (Weighting) 41. Tablets (30% weight): We expect tablets to represent 42%6 of the industry's value share by 2015, and become a key area of focus for PC vendors going forward, 2. Distribution (20% weight): With 351mn PCs (excluding 17mn tablets) sold last year, we think distibution can deliver many benefits like I) improved time to market, i) promotion, and il) each 3. Emerging Market position (18% weight): Exposure to emerging marke is critical, with Units growing 18% over the next five years (versus 10% in developed regions). 4. Brand (15% weight}: In any consumer electronics and enterprise category, strength of brand (as a signal of quality) frequently contributes to a vendor's success. 5. Scale and supply chain (10% weight: Vendors with a higher level of scale and more efficient supply chains enjoy benefits lke ) purchasing power and i) R&D leverage 6. PC product portolio (10% weight): Product success in our view may take a variety of angles including i) portfolia depth and breadth, i) epecs and i) price points offered Vendor Conclusions (Ranking) ‘Apple (71/100—rank #1)—Benefiting from strength in high end and tablet market. In the \raditional PC industry, Apple has adopted a fay concentrated strategy, heavily supolying above the $1,000-plus price point and having a modest global unit and revenue share of 4%19%. Longer term, as the PC market moves to incorporate tablets, we think Apole is bast positioned in this category (scoring @ 10/10) in our scorecard—both in terms of consumer and corporate adoption. This being said, Apple continues to hold a weak ‘emerging market share in PCs and has limited portal breadth in the Mac line-up which ‘may prove as a slight headwind in the near term. HP (70/100—rank #2}—A close second, given lack of tablot vsiiliy. While we do not ddony the strategy HP is taking with webOS on tablets, we simply need to have vsibilty on |) uptake, i) distrbution and ili) ecosystem development before scoring the company higher in’ this catagory. Tablets aside, HP rates wall on our scorecard in most other categories, namely brand and distribution, One weak area includes HP's emerging market position Samsung (58/100—rank #3}—Addrossing most categories well Despite having only 3% revenue and value share in the PC industry, Samsung in our view is well positioned to continue gaining share given strength in i) scale and supply chain, i) brand and ii) porfolo. While the company's tablet strategy remsins pinned to Android (which limits differentiation in our view), equally Samsung was the first of PCismartphone OEMs to launch a tablet called Galaxy (following Apple's (Pad), which has already sold over 2mn unis inthe first 3 months. Lenovo (54/100—rank #4)—Executing well in core areas of strength. Unlike other vendors Which appear more balanced across all categories, Lenovo appears to have various pockets of strength (which fo an extent outweigh the weaknesses). In fact, Lenovo has @ very strong emerging market share (accounting for 30% of PC sales in China, which was. 107Carprrsuss™ “16 March 2011 20% of global PC shipments as a country last year) and a favorable exposure to indirect distrbution channels, Lenovo's weaknesses include lack of tablet strategy, brand and scale, ‘Acer (52/100—rank #5)—Middle of the road. We score Acer a close number five to Lenovo primarily owing to a lower score in emerging market position (4/70). In fact, we ‘observe that Acer infact supplies a strong product portfolio primarily through indirect sales channels despite lower than average scores in brand, scale and emerging market position, Dell (48/100}—rank #8)—Struggling on many fronts. Dell, in our view, is lt positioned to {gain share in the PC industry owing to weak scores in the core areas of tablet (he Streak 7 in our view is not compative), distribution (the company is highly exposed to sell through direet channels) and emerging market postion. This sald, Dell has maintained a strong brand image and given that the company supplies 12% of global PC demand, i stl, has a valuable level of scale relative to peers in the industry, Unit and Value Share—The Story so Far Before diving inta the discussion of our scorecard, we frst think it is helpful to step back and put into context three key trends, we see affecting all vendors in the PC industry: Figure 128: PG Vendors Unit and Value Share Evolution Over the Past Five Years Vendor PC unit share, % Vendor PC revenue share, % 2006 2007 20082009 2010 2006 2007 2008 2009 _2010 HowetPackare 163% 177% 18.1% 191% 170% HowelPacar 161% 183% 199% 1ozk 184% eer 289% 95% 104% 129% 127% Acer 5% 68% 68% 8a% a5% Lanove 72% 76% TSK BO 9.7% Lenovo Tek 80% 79% 82% 91% Torna 40% 41% 48% 80% 54% Toshiba 40% 4a% 59% S4% 5.3% rove 24% 29% 34% 38% AN Apple 47% 50% 73% Bah 89% same a% 10% 12% 21% 22% Samaung 1% 11% 13% 28H 3.2% Sony 10% 18% 20% 19% 23% Sony 25% 27% 30% 27% 32% Fajtsu 35% 32% 27% 17% 10% Fupow 40% 36% 32% 21% 20% others s00% 365% 321% 202% 257% Omer 228% 272% 245% 29% 207% Total 100.0% 100.0% 100.0% 100.0% 100.0% Total 300.0% 100.0% 100.0% 100.0% 100.0% ‘Asian OEMs intop 10 72.8% 248% 27.7% 3.0% 349% Asian OEMs intop 10 23.5% 249% 264% 27.8% 90.5% Unit and value share consolidating for the top 10 PC OEMS. As seen in Figure 128 above, the top 10 PC OEMs accountad for nearly 75% of the industry's unit share in 2010 (80% of value share), up from 60% unit70% value share just five years prior. The largest gainers, over this period have been ASUS, Lenovo, Apple and HP, which have taken share from vendors Ike Dell, Fujtsu and others not mentioned above, PC margins sticky in the low- to midsingle digits. Many PC. companies tend not to report profitability which makes comparisons around profit share dificult to compare accurately. However, based on current information available and our estimates, wa conclude that ‘margins have tended to decline aver time. Longer term, we think execution will be critical as competiive pressures and lower barriers to entry support continued thin profitably Asian vendors are gaining share. Owing to rapid growth in emerging markets (discussed bolow in ‘Emerging Market Postion} and in particular Asia Pacific, Asian PC vendors have gained a significant level of share as seen above in Figure 128. In fact, Asian-based vendors (only in the top 10 PC OEMs) have gained 12% of unit share and 7% of value share in the PC industry over the last fve years rardware 108se morn 2014 so doveeser ess porn see ‘wep lueduog ‘anes aren pe peg iom8 Pm 0H se oped joyeu Bubiowss.JH somyput case YEOH Anan 241 Sa G0! YP %21 sanders Aueduos ot van ue 1 gh ABatexe 241 up yu0p om Siu" J Soe van puodas bea ya YUE! ~ 00 "ua, s20u 94 puvepeou B= ‘ese onoa feu youn desu 2e7) aut uipesigoteaod rEiouo ear pp o) somaya ajay ‘pes Su, vosope ejendioo Shr pear ou 10 sem ~ (GR Hues = ook) oy ow Bulsoiy (ew Hues 5 ° ° 4 4 creed pod oe ‘ ° ' ‘ ° ye ans pue 35 L € 6 e . uoqpod pew Busou © € > , . ae ARTY sor moze wre Kee 1% (ola emus amen a 019 WN Ob z Z 6 9 $ e b 1 'aveys onuansy Od Hue mee 91) RS REZ ELL eu mY loio2 ames oyu oe O19 wot e 9 6 8 z ’ b L 1% ‘aveys yowrew Od Ue, zeae 9% & ow w 2 +6 on ue (porysiom) atoos, oF 6 b (parybrem) yuey ‘284g 2507 0} ff947 nsyng pue ‘i9d ‘SSW HRM 7 SiOUIeD axeYs 1506267 ou) 2g IM Sunswieg ‘gH ‘alddy 51S0bng preddi09g og AuewUUNg “6z| o3nb14 109creprrsusst™ 110 doveeser ess ypo1g ‘yp huwcuog ‘eomN0s 81 spe8u SO eta ma st i Bunauies aes ‘unswies 7 7 oo = = = p Sues pue © © © © y y y y s 9 o (peryBjou) a1095 8 9 proM) YUE, ‘sw30 euoydyews pue oq Jofew Woy, (seyoune7 peysedg Bulpnjoul) seyoune7 jonpoid yaIqeL hey Pue p1eoE109g JOIAeL :0| auNbisCarprrsuss™ “16 March 2011 Tablets (30%)—Competitive and Supply Side Factors ‘A key conclusion discussed in the above section titled ‘PCs disruption coming’ is that we ‘expect tablets to have a significant impact on the PC market and value chain (capturing 42% of the industry's value share by 2015) driven by i) mobilly i) improving affordably, i) effective distribution, iv) significant applications and v) supply side push. To assess who Wins and loses in tablets longer term, we have employed a framework similar to that of the overall PC scorecard. Categories which we use to rank vendors include: i) operating system, i) product porfaiothardware, i) price, W) carer distribution, v) appsiservices and vi) installed base. When judging the Key PC and smariphone vendors in the tablet category, we arrive at thrse main conclusions: ‘Supply Side Considerations \While we fundamentally believe thatthe end dlemand for tablets from a consumer demand perspective could remain robust, we equally believe that the opportunity has not gone unnoticed by mary vendors (mentioned below) that come not only from the moblle device industy but also the tadiional PC market. Before even discussing the strengths and weaknesses of each vendor in the tablet market, perhaps itis worth addressing the ‘demand supply balance risks that could evolve. Here we would note several Key points Significant launches by many vendors. Year to dale, based on our proptetary tablet portfolo database, we believe that some 25 tablets have been launched by the leading ten PC or smartphone OEMs (of which have launched a tablet). Including all other vendors, we count a total of 59 tablet launches to date as shown in Figure 131 Figure 134: We Count a Total of 69 Tablet Launches to Date (25 from Top Handset and PC OEMs) eer 7 usu 7 ‘NetBook Navigator 3 noc 1 He 1 Open Peak 1 eve 2 ows! 1 Panssone a archos 1 Kno 1 Panciata + Azpen 1 Moton Computing 1 Samsung 4 cisco 1 Motorola 1 Tossiva 1 Totar = ‘Source: Company data, Croat Suisse research How will the tablet market play out? With such vibrant competition in the tablet market, we believe that clearly not everyone can win in terms of market share, in particular, tne risk is that the tablet market wil quickly commoditize unless vendors can differentiate in & significant way. We believe thal vandors having a limited competitive edge will suffer from declining GMs and ASP pressure similar to the smariphone segment, wnereas vendors with a diferentiated platorm and ecosystem can extract higher value over ime. ‘Apple (10/10}—A Leadership Position, But How Long Will ItLast? (While certainly not the first to launch @ tablet device, in terms of consumer and corporate rmindshare, we would argue that Apple's iPad has beon the first tablet device to make a significant impact in the market. The question now becomes the sustainabilty around the ‘company’s unit share (which we estimate was ~B5% last year). Longer term, while we estimate share will settle at ~50%, based on our tablet scorecard we think there remains & strong case to be made about Apple's continued dominance in this product segment rardware m1carorrsuss™ tomva sor Operating systoms load. We believe that Apple has been able to effectively leverage the ‘same OS for bath the iPad and iPhone. As shown in Figure 152, while certain features for the iPad are missing such as multitasking, we believe that these are already being or will Continue to be incorporated in future releases. However, we would highlight one advantage versus Microsoft based tablls Is that the company has an optimized touch- screen experience. Clearly continued innovation here is necessary, however, we believe that rival platforms will take sometime to catch up. Figure 132: Apple's 10S Release Schedule Has Been Consistent and Continues to Bring New Features to Market Release Vorsion Features SurOT 1.0) Cover ow, 9) Waual vooomal, i) Phatas,W) Calendar w] SM Messaging, w) Salah wad browser, a) Ren HTN ‘mal, vil) Google Maps, 1) Weigel (Stock, Weather. Caclator, Calendar, x) W-Fi = EDGE Networking Sep07 77) Tunes WF Music store avaiable i) Custom cng tres for $0.89 i) Starbucks and Apolo Tunes W-FI Tunes erinaship,w nvoduced Ped Toucn,v home button doube-ickshontct support for TV out DeeO7 TAS) Googe Maps Locate We", i) Mllge SM Messaging, i) Web Cips features, W) Re-arrangeadie widgets and eane,v) i Free for nar 196 [90 rine), Further contnt management support vi) Sofware Development Kit (SOX), re, W) $19.9 for Pod Toush Jalbe ZO} Support tor WiresohExchanga AciveSyne ad Cisco Pac VPN Push ema unetonaiy i) App Siow.) Suppor for Mabie, ) Mullpte Calendar support, v) Video erentaion expanded, vi YouTube gg’ for Sa Sep08 2.4) Genus plays creation i) Improved iad functionality, Improved Podcast uncionaly,W) Free for Phone users, v) Fre for Pod Touch users wna upgraded to Prone OS 20 Novos 22 |) Ennancoments to Google Maps (Google Sto0t View, Pubic ansit and walking érecons, Adress dplay of dropped Doin, Share locations via oma), i) Severs! Ap Store changes, i) Enbancaments to Saar, W) unos over EOGE ana 56,» Improve Pos unetnaity Jun-08 30) Cut copy, and past, i) Muttmeia messaging including pices, auto, vio fle, Mabie oars "End My Phone" option, W) Shake io shut during plays Pod) Apple push noticaton senvcas, vi) Pas si) Expanded search capabstios Rei 32) nveducoa win Pa) supporto landscape home sero. i)720p HD videos avaiable YouTube App, Mook Jun-10 40) Phone 4 with aver 100 row features, ) Muftaskng, folders, W) etna display integrate, v) mall -unied inbox & ‘treading, vi enhancod camera ané & photo apps (location) anéscape mode, vi) dompererlerpise support vi) Books, 3) Ul eusiemization, x) Sx igi! zoom.) Bing. pear connect, Sep10 ZT} Game conte.) TV show rentals ($0.99), Tunas Png, HOR camera on Phone @, v) HO vide upload ‘YouTube & MabieMe an Phone d ) FocsTine cling am tavortes Novi 42) Main Pad updos, printng i) AePlay (team audi, vioo & photo aver wi, iv) Mullasing Marii__ 43s New Javaserpl engine fr Sela i) Enhancements A:Pay i) Personal holepet supper. Tunes Home sharing Priced for success. Unlike Apple's strategy with the iPhone which was launched at a $600, [ASP and delivered a gross margin of 50%-plus, with its iPad, the company decided for a ‘much more aggressive price point (6499-829 at a GM of 37-42% at fll ramp-up): our potential competiars ar having @ tough tie coming close to Pad pricing even withthe for sale, far loss expensive screens. The \Padincomorates everything we've leamed about buling high value products tom iPhones, Pods and Macs, We eeale cur own AA chip ur own software, ur ‘own battery chemistry, our own enclosure, our own eventing, and tis result in an needs product ta. geat pce. The proof fs wil bn the pricing of eur competitor’ roducts which wl kay oor Joss for mor." [.] “wee priced, ‘Pad prety aggressively, so wore out fo win this one.” Vertcaly integrated. We believe that in the near term, Apple is advantaged in the tablet market versus peers given the company’s vertically integrated strategy; the Ad processor Is designed in house and the 10S operating system is maintainad in a closed environment However, a counter argument can be made as to whether open systems and multiple vendors will prevail in the longer term (Android of Windows Phone 7 plus choice of chipset provider ie., Qualcomm, Nvidia, Texas Instruments, Marvell et.) Leveraging tho existing installed base, One key advantage of the iPad versus peer offerings is how it works with existing Apple products and services. For instance, a user can leverage existing ‘Tunes content (ie., music and select apps) and contacts on their iPad out of the box. We think further methods of synchronization, whether on Apple TV, iPods and new product categories, only enhances the ullity of the iPad, rardware neCarprrsuss™ Carter distribution. Given tablets are ultra portable devices that are increasingly soling With cellular data plans, we think dlstibuton will naturally evolve toward the carer channel. While many smartphone vendors (for example, RIM, Samsung and Motorola) are well positioned to take advantage of existing relationships, we show in Figure 133 that [Apple has the widest distribution across the top 30 mobile operators globally in terms of subscriber base relative to other top 10 PC OEMs. Figuro 133: Top 10 PC OEM Carrier Distribution—Samsung and Apple Lead the Pack, Given Their Existing Mobi Source: Company dete, Crest Suse research Applications from the get go. Apple enjoys touting the number of applications available for the iPhone at over 325,000 and as shown in Figure 134, Android is a not so distant second with ever 200,000 apps, While the number of apps does not directly translate to a vendor's advantage in terms of product success, we do think it has proven to be an Important contributing factor. In fact, there are already ~65,000 apps available forthe iPad alone, merely 8 months afer launch 113Carprrsuss™ “16 March 2011 Figure 134: Apple Leads Other Mobile OS in Terms of Application Ecosystem Momentum and Media Offering Featurel Metric ‘Apple ‘Google Nokia RIM MisrosoR Number. 25,000 (85,0001Ped)>200,000 ~30,000 “20,000 “6,000 Book store ocksire Amazon's Knale x x x Music! Video content Tunes Amazon MPS Musi Stove Taligis! Zune Marketplace Muse catalogue size (mn) >13m0 Pram 2900 nn em Movi catalogue size (0005) oss x x yes ys TW show eaalogue size (0005) 355 x x vee yes Developer communty Number of downloads (rn) >10,000 >2.500 1500 Na Na Revenue sharing wih developers 70% 0% 70% 10% 70% ‘Scwce: Company data, Creat Suisse research ‘Android Vendors Collectively Will Win—Howover Which Will Prevail? \We count that Android is being shipped on ~85% of the total tablets selling or soon launching from the top 10 global PC OEMs. As a platform, Android has proven itself in the mobile landscape, now accounting for 31% of the world's smartphone shipments in Q410. While we expect the platform to evolve quickly aver the next year, we think it alows PC land smartphone vendors alike (without a proprietary OS solution) to design competitive products quickly. Longer term, however we question the level of value that each Android vendor wil collectively attain given the OS essentially puts each hardware vendor on level ground Rapid OS innovation. As shown in Figure 135, Android has demonstrated strong progress ‘ver the last two and a half years on the mobile side, just recently evolving to embrace the tablet form factor with the announcement of Android 3.0 Honeycomb OS, Figure 135: Android OS Has Shown Significant and Consistent Re 108 Since Its Debut in September 2008 ‘Sep08-reb09 10/11) Maps ad dolaam i) Suppor for sang atachmert or NS, ) Suppor formarquce ih ayOWe peo 115) Satoh cor vdeo, I) Upgraded sok koybosrttaxtpodictin, i) New widgets and fia for Home seroon Seno Ti) improved Ando Hark, i) Universal sear, i) Support for COMAIEVOO, VPNS, ello speech, 1) Fee tn by tum Googie navigation 008.09 2012.1) Optimized hardware spe6d, i) Updated Ui Improved maps. exchange intaratin, Live walpapers ay-10 22) USB tethering») Fash 10., i) Pectormance optimization, ntgraton of Chrome's VB JavaSerot engine for eet "25/2. Revised Ul design, i) Support for lrger sree. i) New 30 game support Feoa0 {30 1) Optinization fr tablets, I) 30 desktop with widgos, i) Refined mut-askng,W) Browser enhancements 6 ‘Source: Company dal, Crt Suisse ros0erch Applications—quickly growing. While Android remains @ second to Apple's App Store in terms of total apps offered as demonstrated in Figure 134 (Android at 200,000-plus versus ‘Apple at 325,000-plus), th platform is quickly catching up. Wider Google strategy wil benefit the platform. While we think Google fas itl interest in becoming a hardware OEM, Android remains a key focus for the company to drive mobile search and hence display revenue in the future. In fact, Google's strategy has now ‘expanded beyond the PC and mobile phones to now include tablets and (potentially) TVs. Android and the ecosystem. Discussing the prospect of every vendor that supports ‘Android is beyond the scope of this report, and as such we only discuss key OEMs rolevant to our scorecard below. This being said, it is important to recognize that collectively many smaller vendors are gaining relevance in both mobile and tablet markets, by using Android, and these vendors too will influence the ultimate direction, scale and brand of the platform: rardware "4Carprrsuss™ Samsung (6/10)—extending the Galaxy brand. So far the Galaxy Tab (see Figure 130 above for specifcations) has already sold some 2mn units (end-January), which we think represents strong sales despite being only a 7" tablet and runing the non-tablet optimizes version of Android (Froyo v2.2). Utimately, we think ths lays testament tothe strength of, the company’s brand, scale and distribution. Given Samsung's heritage in both moblle devices and PCs, we expect the company to become a major player in the tablet market longer term. Though Windows Phone 7 and a corresponding tablet version of the OS may bo on the plate for later this year, we thik Android will continue to be the dominant platform for the company. (Its unclear f the company will extend the Bada OS to support, tablet form factors inthe future depending on how viable the OS is on smartphones) Motorola—one tablet so far, smaller form factors on the way. The XOOM is Motorola's first ‘Android 3.0 Heneycomb-based tablet launched last month with Verizon in the U.S. The tablet features 10.1° display, Nvidia Tegra 2 dual-core processor giving 2GHz of processing power, battery life which supports up to 10 hours of video playback and 32GB. fon Board memory along with GB of RAM. The device also comes with both front facing and back facing cameras of 2MP and SMP respectively along with support for Google Maps 5.0, access to over 3mn Google eBooks and Google Talk for video and voice chats While the price point of $799 is significantly more expensive than the low-end WiFi only iPad at $499, we expect more affordable options from the company in the future at smaller screen sizes (7" tablets have already been discussed by management) Dell (5/10}—two tablets so far, @ 5° and 7° Streak. It appears that Dell is committed to launching tablets in the coming months based on Android givan the company's recent ‘Streak launches (7" more recently announced at MWC). Given Dells heritage in PCs, we believe a modest market share in the corporate and consumer tablet markets long-term is, not out of the question. This being said, itis dificult to assess success at this point given a few products and limited commentary tram company management on is long term tablet strategy. ASUS (4/10}—Android and Windows 7. At CES this year, ASUS announced the avalabily of four new tablet devices (Slider, Eee Pad Transformer, Eee Slate EP121 and Eve Pad MeMo), three are based on Android (2.2) and one on Windows 7. While we are Impressed with the breadth in product form factor (each appears to serve @ unique use cease), we think the company may be fighting an uphill battle longer term given limited carrier distribution (currently), and potential afficulty in supporting the matile supply chain {new chipset vendors, OS suppliers, et) in addition to the company’s current PC supplies. HTC—only one tablet, but sti in experimental phase. HTC has been relatively later than other Android smartphone vendors (like Motorola and Samsung) in joining the tablet bandwagon. However, the company announced its frst tablet device called Flyer (7" tablet based on Android 2.4 Gingerbread) st MWC in February this year, which is expected to start shipping in Q211. With growing brand recognition and strong high-end smartphone distribution, we expect HTC to benefit from tablets long-term HP—A Shot at Differentiation with webOS. Following HP's acquisition of Palm for $1.2bn in July 2010 (announced April 2010), the company has just recently announced its first tablet device, TouchPad, on webOS 3.0, ich is expected to stat shipping summer 2011. While much remains to be seen in trms Of strategic direction for the platform and supporting hardwarelintegration with HP's broad porfolo, we think this represents @ good opportunity for the company to differentiate versus Android longer term. As shown in Figuro 130, the TouchPad features high-end specs lke @ 1.2GHz Qualcomm Snapdragon processor, 9.7" screen and is expected to ‘come in 16GBI92GB flavors, RIM—How Much Potential Does the Playbook Have RIM entered the tablet market with the Playbook device, first announced in September 2010, for a March 2011 launch. Intally, the company launched only the WiFi version but 6carorrsuss™ recently at MWG, it also announced its plans to add other connectivity options like HSPA, LTE and WIMAX, which will be launched in second half of 2011 QNX is @ differentiated OS. While itis hard to imagine yet another mobile OS in an already crowed landscape, we think RIM's NX OS makes sense for the company as itis & significant improvement over the existing BlackBerry OS. Addtionally, QNX brings improves functionality to the mobile environment in terms of real-time processing and utta-low latency. In fact, the OS has traditionally been used in a variety of vertcals— ranging from telcos and biomedical to automotive and gaming. Finally, we would note that QNX is a true multtasking OS, unlike peer offerings which tend to pause applications running in the background Verticaly integrated. Similar to Apple, RIM is @ vertically integrated company in terms of hardware and software control. With this sai, RIM is relying on Texas Instruments to ‘supply the application processor forthe Playbook, wich at this point has no disadvantage (0 peer tablets running on dual core GHz processors. Rikt's NOC based architecture could provide some benefis. Here management has argued that RIM's proprietary Network Operations Center (NOC) could further differentiate ne company’s products in the tablet market by providing compression and limiting the download of data from the server (given this is push technology) and thereby improving battey life. We think additional benefits from RIM's NOC could stem from improved security and new opportunities in e-commerce. Carter distribution channel. While we do not include RIM in Figure “33 (given the company is not a top 10 PC OEM), we would note that we expect the company’s distrbution to be similar to what is shown for Agple. Given our view that tablets wil Increasingly evolve toward being distrbuted vis-A-vis the carrier channel, we think previous mobile operator relationships RIM has established in the smartphone market wil, prove a valuable asset in the tablet market. ‘Apps strategy stil in limbo, As demonstrated in Figure 134, RIM's application strategy has. failed to pique developers’ interest with only some 20,000 apps currently (versus Apple and Android at over 325,000 and 200,000 respectively). While it remains unclear what the company’s strategy willbe in the tablet market, we think the GNX OS may provide an ‘opportunity for the company to break its current trajectory. 116creprrsusst™ ‘sjeuseyo pap pue o2upur Woq ul sexeys yiaK0 mo} anasqo aM “saynqujuco Sueciuno yoes seed au ow uede ‘vayo29 Si) Jt pauogpsod jm aynb swwodde .SvA\i0, 2081052 U) yOnowty :(0472) 8124 (x) }2aupusnszan (aveys agep) UAB sty ot “spout Jounsu0o ey oexp ® Uo Kee Suro Ut ssovBoid Supe u2eq 824 fea a4 L0H) 9a ‘aueys lauueyojoaup snssan oq Aisepow © 0} Bulmo pUE (ge) axpur snssan (areus yee) sjauseyp yoou0 1 pasodse s10w suiewas afddy :LoL/s) dy He 3 Kyunyodéo squewBes yynaib 10464 24) 0} pata 99 0} seadde 104 se09 Aueduoo 2a ‘voInauysip jeupul Sue} UL Yoi9) Yau UL avEYS J2yBU ‘AuBis 8 sey nq 'oKNqUip jo suey UF eaIYsO, Se aijoid JeIus © SeU 0n0U8) :(0V9) onoUST 962 Aluo Snsian sjauueyo Uoangsh } Ur aieys 99 argerone © sey eqIYsOL Tego 94g © Buinay fluo aydseq ove) ‘souueyo uonnquasip ‘uy 40 yoee uywa sevoys peduElER- jam ene _yoreosos ss ypero ep Aveduog YoueD -mn0g an KioBa}eo uoee ul Oy) UONnAUIIP poaupur ula sey ye uoK'sod enero: ‘IYS SIOPUEA PUE SDq 10} [OUUEYD Aq UOANGINS|G JO WONTON oe aIn14 "7Carprrsuss™ Distribution (20%)—Exposure to Favorable Channels Given our view that PC volume (excluding tablets) will rse to 392mn/414mn units in 201212015, up from 351mn last year, we believe distribution and supply chain will remain & critical factor for vendor success in the industry. Here we expect that vendors having an extensive reach and a well organized strategy in terms of channel (indirect versus indirect) can teap multiple benefits. (For details on our rankings please see Figure 138.) We would note that @ strong level of distribution affords vendors multiple benefits including the fallowing = Improved time to market. While this may seem obvious, taking too long in getting new products tothe market is fraught with risks and may lead to disappointing sales. "= Shelf space, Trust earned with retalers over the long term may lead to higher promotion for products, with price protection and minimal price volatility © Increased penetration. We think the right channel partners may allow vendors to increase penetration in keylgrowing regions — for instance in emerging markets we think exposure to cartier channels may improve a vendor's tablet sales/ positioning. In order to actually measure these factors, we have locked at a combination of indirect channel distribution (favorable) versus direct distribution market share (less favorable) and cartir relationships (previously discussed in tablets section above, soe Figure 133) HP (9/10}—Broadest Reach and Most Favorable Exposure Given HP's global scale (see discussion in the below sections), i comes as no surprise that the company also executes well in terms of distrbution, As seen in Figure 135, HP has @ high indirect distribution share of 19%, which we think offsets @ 13% share in direct, dlstrbuton (keap in mind indirect channels account for some 5x tha number of units as i direct channel) ‘Acer (9/10}—Strong Exposure to the Indirect Channel [Acer ranks just as strongly as HP in our view given a strong indirect channel share (16%) with even higher shares in the faster growing categories of general merchandisers and consumer electrons stores within the inaiect channel Apple (5/10)—Improving Distribution, But Unique Headwinds Unike all other vendors (except for Dell which we discuss below), Apple has a higher share of unit sales through direct versus indirect channels (8% versus 3%). Ths is largely owing to the company's 10% market share in the ‘Direct Fax/PhonelWeb and 7% market share in ‘Direct Salesforce’ which includes the company's web store and retail locations. This sad, this stategy continues to work well fr the company versus the move away from this approach for other vendors in the industy. Carror channel advantage vorsus PC poors. AS we have mentioned above in the tablet scorecard section above (Figure 135), Apple also has a strong positon in the cartier channel by virtue of the company’s prior mobile relationships. While this is not yet & channel for distribution for acitional PCs, i cannot be ruled out longer term as devices become increasinaly mobile and include cellular connectivity Doll (2/10}—Move Out of the Direct Medel Still Under Way What was once Dells strength, the direct to consumer model, might now be the ‘company’s Achilles heel. While the company is transitioning away from this model, they hhave a 43% share in the direct channel versus 4% in the Indirect channel and for this, reason, we score the company lowast on our scoracard in terms of distribution 118creprrsusst™ _sejewse 53s ppe19 Rp huediog 2020S 290 pue 190 InOge HuKH BNA HOLIP) 89H (946118 onous7 se ub ‘5 Auwau aveus © yin) a18us (0 sjeyrew Budojerap ur p=uoMh “sareys jayeu SuSiewe oy81y arey siopuan sey ie slepuan Pevoqueweloje ely Woy epsy (Aienpoedse: Se 1%9) SKEW padojenep pue Cuxo}srap 94) oq! o22ys sew paovepeG oncy Cunswes pur snsy weg ‘(one) Sunswes isnsv ‘slopuen od fue 9 uoxowebjewe Ue S11) ueN® onousT se ANB se 10440, ‘2! iuop OM HPL /mK6e) eIES Pedownep ‘suofia! Sufvewe Buowe 1se18e) SuNo16 ‘18 you ‘Souuned oyped esy Burdopnap aup eovoury URE? pue oye eisy "1 squauibes Bumoi Jase Ul seus 19L61y WA siopUOA Jone} om A0BoqO UDED Slayew Gudojarap pue sayew oyu SJEpUBA Oe 21! 01 ‘ABojopOUFoU pleDd100g (2 st fa eae) syn 4 ego 00 Dal Tahu Sub ious "80K yo] 9 sy (WEHIeU! padojnep 19) YOWO %9 © Sn8!eN) %LL~ 19 BOYD © 18 IDK 94 80] 00 OuEODUY sq Sey s}ueW GuBreWs ul puewep ‘Da wnjuowow BuureS sjoyew BuBs0w (1 ‘@2LUS UIeD 0} pouoR|sog ys0q Bunsiweg pue ‘snsy ‘onoUR” YIM 'S1EDA OAl4 ISed 199 PeyElo|e00y SeH SjOyJeW BUIBIOWN3 Jo BoUEAA|eY 1c} 21ND 19Carprrsuss™ “16 March 2011 EM Position (15%)—Exposure to faster Growth Over the next five years, we expect PC shipments in emerging markets will grow at 18% ‘compared to only 10% in developed regions (as shown in Figure 38). All this suggests, that emerging markets will represent nearly 55% of global PC units (excluding tables, this will be as high as 64% by 2015) and ae euch, we highlight this should represent a core focus for PC vendors looking to expand market share. Based on current regional market share data, which we detail in Figure 137, we reach the follawing conclusions for each of the vendors. Figueo 138: We Expect EM Units to Reprosent Nearly 60% of Total Volume Long Term, Up from 45% Last Year ‘GAGR GAGR Units in mn 2002 2003 2004 2005 2006 2007 2008 2009 2010 20116 20126 20158 CAGE CAR Deverped marketunts 10h tie tae tar ta) —182 tea a70 00 2ae TONDO aH Emerging marketunis _&& 5059749112129 139 169_—*197_298_— 3037.84 10S Global unite We 16448321223 6402-908 368 ASD SOB TINT 14% Y-dovolopod market 70% 69% 68% 85H GIN STH 58% SSK HAH SAH 5IH 45% Shemerging market 50% 31% 32% 358 S054 494 aa 45H 40% OH 47H 55H “Sours: Garter, Credt Suisse estimates Lenovo (10/10)—Leader in Emerging Markets Lenovo currently has a 10% global market share, but of this share the company leverages 2 high 15% emerging market share and only a 5% developed market share. Furtnermore, Lenovo remains highest exposed to the developing Asia Pacific countries, which are ‘growing fastest among emerging regions. HP (5/10}—Strong EM Share, But Equally Highly Levered to the Developed Market While HP does have a strong 14% emerging market share (second behind Lenovo), we would point out that the company stil has a 22% share in developed markets Furthermore, it may take time forthe company to regain its footing in China (19% of global PC volume last year fllowing the company's early-2009 hiccup (warranty not honored for faulty systems) and subsequent share loss as shown below in Figure 138, Figueo 139: Over Past Four Quarters, ‘= Markot Share in China Has Nei ly Halved, Primarily Benefiting Lenovo. ‘hina marie aio aay asa aaosattoazie sto nto Teno Tsk 258% 24 —-aTSK 226% —2BaM BOK Oa Howiot-Packard Mah 126% 53% TTT TOIH 79% BOK, Founder Electrons 10% 66% THK OTK TO TKO Top Svendors'share S234 658% aT —~Su—~AOM—~C«ST «STON =A “Source: Garner, Cred Suse research Doll (4/10}—Developed Market Share Larger than Emerging Market Share We believe Del's market share position by developed and emerging regions resembles nat of HP, to only ata slightly lower level. For this reason, we give Dell @ lower score than 1P. (On the positive, Dell did not have similar issues to HP, but at the same time, the ‘company’s share in China has not shown any meaningful growth over the last two years). ‘Apple (2/10)—Weak Emerging Market Position While Apple is increasingly looking to introduce products tke IPhone and iPad into femerging markets, the Mac business we think remains at a disadvantage given its high {$1,000-plus price point. For this reason, Apple's emarging market share is only 1% today, versus 7% in developed markets. Over time, we do not expect a radical shift strategy with the Mac product line that would ater this allocation of share and for this reason, we score Apple lowest on our PC scorecard in the emerging markets category. rardware 120Carprrsuss™ “16 March 2011 Brand (15%)—PC Industry Is No Exception ‘One metric which has consistently proven itself as an important contributor to a company’s success across nearly all industries is brand, and the PC industry is no exception. To fassess this metric, we have summarizes the brand rankings of Millward Brown and Interbrand as shown in Figure *40 and Figure 144 Figure 140: Millward Brown’s Brand Rankings Figure 141: Interbrand's Brand Rankings “Ha BE “Source: Milward Brown, Cred Sulss research “Source: horbrand, Cred Suisse research Apple (10/10}—Synonymous With Quality? Based partly on Apple's wider product strategy, we note the company has consistently risen up the rankings for brand, now being regarded as the third highest ranked brand Worldwide according to Milward Brown and seventeenth based on the Interbrand assessment. We believe this stems ‘rom a focused product strategy in most ters where quality is emphasized over volume. of Market Ca Past Five Figuee 142: Brand Value as a Pe RRRGE ERERE GREE RRERE ERERE ERERECarprrsuss™ \We belave itis this strategy combined with strong brand recogrition which allows Apple to quickly address new market opportunities, ike the tablet market with the iPad most recently. Furthermore, a strong brand has put Apple in a solid position to efficiently resolve productrelated issues without significant consumer backlash—the most recent being IPhone 4's antennaelreception troubles which surfaced in July 2010. Taking into Consideration these factors, we have rated Apple 10/10 HP (10/10)—Rising Up the Rankings Inine with HP's continuous PC market share gains in recent years, the company has also been improving its brand ranking in each Millward Brown (12° globally in 2010) and Interbrand (10" globally) over the past five years, Accordingly, the brand value now represents close to 31% of the company’s market cap, up from 20% only fe years ago. \We think that a strong global brand positions HP well to gain share as PCs continue to penetrate emerging markets. Last, we would note that given the company's acquisition of Palm last year, we think ths presents an enormous opportunity to leverage the company’s brand across new product segments lke smartphone and eventually tablets that are based fn its webOS platform Doll (7/10}—Significant Slip in Recent Years Despite ranking in the mid-20s for global brand rankings in 2006, Dell's ranking has fallen ta 65/41" place according to Milward Brown and interbrand respectively as of 2010. We believe this deterioration in brand value reflects the company’s declining market share in the PC industry and the gradual erosion of the direct to consumer businass model. Yet despite ranking much lower versus history, it appears brand value (on average between Millward Brown and Interbrand) is a staggering 40% of the company’s market cap. All this Said, Doll stil ranks inthe top 100 of global brands, which stil represents significant value Scale and Supply Chain (10%)—Critical for Leverage In 8 commodity business like the traditional PC industry where levels of product and technology differentiation currently are limited, we believe that scale remains of eriical Importance delivering a few potential benefits Purchasing power. For key components, the ability to bulk purchase, achieve economies of scale and procure components at times of tighiness in supply chain, all can become key competitive advantages in torms of time to market and cost within the industry Figure 143: Purchasing Power in Terms of COGS for the Top 10 PC OEMs ‘ope Computer Tostiva (a) esr Group Fujitsu) Sony (6) ‘Souret Company dala, Crea Sues eatates (o) neues mobile communication, dita media networ, and PCsnetwork segments (2) Includes PCs, networking, and egal maging seamen {Gi notes syatoms, network, PC and moble segments {@ Inckides nondsote networks, and PCs sogments 122Carprrsuss™ RAD leverage. From an R&D perspective in both absolute terms and the abil to leverage ‘across platforms and product sets, this can become a key compettive advantage within the PC supply chain Figure 144s Absolute RAD Spent Companywide for the Top 10 Global PC Vendors USS in milions unas ethornize sated Sy ORCC Samsung 3518 3.683 3041 3402 am Toss 387 3007 3.981 3.930 3se Fujitsu 2osr 282 2500 2630 2407 et 463 498 610 609 67 asus zat 231 290 405 511 ‘Sowee: Company data, Great Suisse research \While in our scorecard we rat sach PC vendor on various success attributes, we think the category of scale should be measured across all vendors’ relevant product categories, not Just for the product being discussed. Taking this into account, we would score the major PC vendors as follows, Apple (10/10}—Focused Approach Around Component Rouse and Manufacturing [As shown previously in Figure 143, Apple products account for nearly 15% of all core hardware component purchases (of tha top 10 PC OEMs) despite having a much lower share In all end markets. (Apple's global PC and smartphone unit shares wore 4%/16% last year). Component re-use. As we demonstrate in Figurs 145, one of Apple's strengths is its ability to leverage a large number of similar components across its product range. Infact, the top 10 shared (value-based) components amount to some 38%/49% of the total BOM of each tne iPad (8G + WiFi 16GB) and iPhone 4 (32G8), respectively. We think this approach to the industry is unique and also allows for the scale effect for Apple to be magnified—for Instance, when Apple is placing orders for an Infineon chip, the company's entire product, line can be leveraged to bring total cost down. Last, paring down the total number of unique components provides a level of efficiency in working with particular products. and supplies, whict is not necessarily the case across the industry ‘Strategic sourcing, Last, is important to mention Apple's forward thinking when it comes te product portfolio and component needs. During the company’s F1Q11 conference call, (COO Tim Cook spoke about the agreement Apple had made with @ number of NAND flash suppliers in 2005, “on th operational ste of the house, 28 you probably remem, we have historcaly entre into cortin agreements wit front people to secure supply and chor bonefts. And the largost ene in tesied over $1 bitlon. Because we antpated that Flash would bocome increesngy important across cur entire produc line, and increasingly mportant othe industy, and so we wanted 2 secure supaly forthe Company, [.] And wo constant lok fr more of thes, So the past sevoral quarters we pening comments. ~ Tim Cook, Apple, COO 123creprrsusst™ ‘Yoreosas essing ypaiy wabypaueg e2mes :* vy Budd! +E Ped! _eionpuoatues Borea ‘erkg (@ wospenis (poo (9 Cunswes rey (@ Bunswies (e204 nctns Wabyeyog ame TouoTgT puE eR Ta pore WOE "poo Sep FEL {3809 wevodwie> oF TOIT TET sao-v duo-reuS smsosues YOO NSO por woo} uegoseg YAGO-M soesaaoig suanesdY WY used ONYN OW! 89 Ze Alowsy SuD-RIK ata museronono anapedeD “hangzedsos ouoyd! Josseooid puegaseg '(Bunsweg woy) Kiowew QNYN (Bunsuegraiddy ws +96 Ped! PuE F OUI! O19 ‘Ss9jo\RHOAONY sqHOUOduIO® oj0:08Ip Bulog fofew ‘suouodwoo 266 J0 dh opew 61 ¥ eud4.s! GD ZE eu “SIoNpod ssO1De squauodUICD fay Jo BuLeYs auf YBNOAIp 1oMod Bi Jossooais suoresndy Ma zx wseld aNYN OTW 80 @- Kowsw cua sn poy uaansusnoy oweredeo TRE TTT Ta Tou Seo duoabus SdO-V dug abu (6 voaoosuess Sunn oniostued SN (2 (@ wossesoia suonesieay wey (@) Rowan ONY (Z pan wares wrod IM + DE Ped! 84140} WOR 8MR Jo 2464/48 10} 1UNOD>e syuoUOdoD pareys o} dor 2501p ‘pOUIqUIOD ‘8 [2PoU! veesosyono1 pu (woopeoig fa) preog YRooteng/L AIM *(voeuyLy] oS) sosseooid py eR Bupnjou sia na ‘swwouodwoo E09" ns sueUodwos 5} ued 245 41M PU LIM + 8 Ped! “BO 91 949 0 posedusod ‘Sojouotoiya wleua Aiddng pue Buiseyoing saBeuerpy YoIUM de8AO uauOdWOD JUEOMIUBIS WHIM SHonPOId SUBISeG aiddy :sp1 ounb14 124Carprrsuss™ HP (8/10}—Largest Scale Across Vendors Whether it is owing to HP's leading 18% market share in the PC industry, or more importantly from a purchasing power perspective, the company has material advantages from its scale, in fact, as shown in Figure “43 that HP COGS accounts for as much as 23% of the market for major components (ofthe top 10 global PC OEMs). ‘Samsung (8/10)—Product Lines Similar to Those of Apple Given Samsung's wide product breadth in the consumer electronics market, it comes as no surprise that the company accounts for nearly 18% of core components purchased (of ne top 10 PC OEMs), second ater HP. As the smartphone and tablet markets continue to ‘expand and gain more relevance in the global consumer electronic market, Samsung also appears well positioned to gain further purchasing scale and efficiencies. Last, we would rote that Samsung is in unique position to potentially grow scale faster than competitors, given the company manufactures its own processor (Hummingbira), memory modules and LCD technologies (fo name a few) which are also sold to its peers. Doll (7/10}—Highly Levored to PCs, New Markets May Help Build Scale Dall is among the top five PC vendors with a market share of 12%, and we estimate the vendor accounts for some 13% of core electronic components (shown in Figure 143, ofthe top 10 PC OEMs). While the company has clearly begun inroads in the smartphone ‘market (with the Streak device) and tablet market (with larger versions of the Streak), we think any scale the company gains as thay expand into these new product categores will be gradual 128creprrsusst™ 126 doveeser ess ypo1g ‘yp huwcuog ‘eomN0s '000'1$< Susy Buberem ay puwneing enyp ssize poyecnces hepin wou om 82] PUP df By oe Kg, ig ue ey —porayo ane yo nen pu (0) sy ra "peeds Pop seared 2) ey (Oyu) wos “Bry 7nd Le “a meray yin poe mew om stu "ugefe Bun e760 (8 ‘20h80068~sosnag sous yooqyabsine ws agpocuo> siopuse go ecu sao psig a Bueesppe eq ct ide wom oH pasn sol pve Burau. ru aM up wershs Bune ce ws & ax smopuny 'SO 1 sopu un pot Sypoqiu fo (608) \apeen roped | 9 ay fog pu ang qn 06d snsin syeomou jo (se 08) arsed Wau jxBie a Say SRV THIOL TIOOHON feceecece egecnes ep eeceeced ‘syOOqaION pue SyooqIeN 40} BSeqEIEG O1o}MOd JONPOd Dd 90 aunbisCarprrsuss™ Product Portfolio (10%)—Balanced Portfolio Is Key In a PC market which is constantly changing, we believe @ well balanced and diversified product portfolio offers a vendor a strong means of differentiation, For this reason, we do not think @ one-size-fts-all approach works, but rather one which works across many form factors, product types, specifications and price poinis. Based on our PC product portfolio database chown in F gure 145, we attempt to take these factor into consideration, among ‘others, as shown below in our PC scorecard. As such, we score the vendors as follows ASUS (8/10}—wide variety of product SKUs, We observe that ASUS has among the highest number of product SKUs each in the notebook (103 products) and netbook (87 products) categories. In fact, this allows the company to leverage higher scores on mast other categories in our scorecard based on this product breadth, an example being price point. One area where ASUS tends to score lower Is in core specs, including Weight, battery life and CPU. Acer (8/10}—similar to ASUS product portfolio. While the Acer product portfolio is not as extensive as thal of ASUS, the company tends to score higher in the product spec categories for each notebook and netbook, in particular CPU LHP (8/10}—robust portfolio for largest PC vendor. We noticed that while the total of HP's product SKUs number i¢ much less than ASUS and Acer, the company stil ig able to design a portfolio of devices which appear average across most mettcs including weight, battory life, touch-enablad and OS. It's worth nating that in the notebook category HP's 19, products are equally spread across the major price ters, which we think is a product of the ‘company's diversified distribution (to both developed and emerging regions) Doll (7/10}—middle of the road. We find of the top 10 global PC OEMS, Dell offers the fewest products to choose from (16 notabooks and 2 netbooks). While we think this may bbe a function of the company’s high global commercial share (see Figure 147) relative to ‘consumer share, However, longer term, if the company looks to increase penetration in the emerging markets, we think broader array inthe portfolio is necessary erates 77h 179% s89% 181% Apple (4/10}—targeted strategy continues for now. Similar to neatly all of the company’s product ines, Apple adopted a relatively focused approach to the Mac business in terms of total number of product SKUs and the range of specifications among these devices. urthermore, we would note that Apple does not compete in the netbook market, and Instead offers a $1,000-plus alternative (in the notebook market) called MacBook Ar. For nese reasons, we tend to score Apple higher in the categories of weight and battery lite, but lower in trms of screen size and price point. v7Carprrsuss™ Smartphone Growth to Continue It is no secret that smartphone adoption will continue to increase at the global level ana drive robust volume growth forthe industry. This, in our view, is driven by a combination of factors lke carrier push, handset and chipset vendors driving smartphones to lower price points, and increasing consumer demand pull. We believe that most of the long-term forecasts for smartphone volumes are based on an estimate of smartphones increasing a& ‘a percentage of overall mobile handset market. However, given that smartphones continue ta be significantly more expensive than a low-end device, we believe that any long-term forecast has to be based on linking the income distribution and total cost of ownership to device affordabilty. As such, we have used our proprietary analysis to look at the smartphone market based on three different approaches, "© First, we have looked at the smartphone market from the perspective of atfordailty, taking in to account the total cast of ownership of device whichis often overlooked. © Second, we have looked at the smartphone market opportunity by looking at how smartphones could cannibalize adjacent consumer electronics markets. "= Third, we have analyzed the overall mobile phone market by price bands to determine the price tiers that smariphones can realistically target aver the next five years. This approach relies heavily on our proprietary bill of materials (BOM) analysis, Alfordabiliy/TCO approach supports a long-torm addrossable market of 2bn for ‘smartphones. Based on out proprietary model that takes into account the total cost of ‘ownership (TCO) for @ smartphone, income distribution, and penetration of the addressable market, we conclude that the addressable market for smariphones could be {as high as 1.96bn longer term, We define TCO as the upfront cost that a consumer pays for @ smartphone combined with the annual service cost for @ basic voice and data plan associated with that device. Our smartphone model suggests that by 2018, the global ‘smartphone subscriber base will reach 1.92bn, Le., 98% of the 1.96bn addressable market Based on this long-term estimate, we believe that smartphone volumes will grow from 297mn in 2010 to 594mm O4bn in 2012/2018, implying a CAGR of 28.5% over the next fove years. Cannibalization of several markets simultaneously. Recognizing that the TCO approach is only one way to model the smartphone market, we have sanity checked our estimates by also locking at the incremental opportunity arising from smartphones cannibalizig other Consumer electronic devices such as portable media players, PNDs, digital sill cameras, portable gaming devices and midend phones. We conclude that cannibalizaton of these segments supports a smartphone volume estimate of 988mm longer term, which suggests, nat our 2075 estimate of 1.04bn units based on affordabilty is achievable BOM oxtrapolation and price point penetration analysis also suggests 1.04bn smartphone units by 2015. The third and final approach we have considered involves analysis of smartphone price declines over the next five years and evaluates how quickly smartphones could penetrate lower price tiers. Indeed, based on a teardown of the $800, XpressMusic from Nokia, which we would define as a lower end smartphone, we conclude nat through improvements in chipset efficiency and scale, the ASP ofthis device could fall to as low as $112 by 2018 compared to $300 in 2008, All his while stil keeping the core functionality of the device intact. Thi, in tum, leads us to belove that penetration of lower price points could conceivably result the smartphone market growing to 7.04bn units by 2015, which is in-line with our current estimate based on affordabilty analysis, 128carprrsuiss™ semaren 208 Figure 148: Smartphone Market—Long- rm Forecast Based on Three Different Methodologies Analysis Methodology sult Implication [ou Cantal Propratary Codi Suave ode a1 bn adrassatia aka Tar By BUTS, wa suivala 1S8bn peopl Tobe ‘Ownership forecasts the smartphone market using the smartphones by 2015, up fom potential smartghano suBscbers, and wo aro TCO (otal cost of ownership) ofa 7b. in 20082010. cureny assuming 1-820n amargnone smarghone, income distbuton and ‘ubscribers, which mpl about 96% of te Ponavaton of tho accrossanie markt ‘accressabia markt Henos, we Beeve our fmarohone subscriber esti are achinvabie CCennibalzaton of Forecastng smartphone markt based on 988mm smartphone unitsin Our .Odbn estimate or smarpnone volumes By CCE devices incremental opportunity lo cannzalize ather20%S based on smarphanes 2015 looks achievable a8 ean aio Bo: Consumer slocronc devise suchas MP3. cannibalzing othr CE device our cannibalzaton wor whin sugge Players. gaming corsous, NDs, cameras segments Smarghone volumes could be as high as 988m" and mid-ond phones 150M reduction Looking a! amariphane market and pcs band bilo ma cou po ‘Soure: Great Sse reson? TCO versus Cost at the Point of Purchase Debate We believe that one of the key determinants in trying to estimate the size of the smartphone market longer term is affordability. To determine the atfordabilty issue, we have looked atthe total cost of ownership (TCO), which is the cost of smartphone device ‘along wih a voiceidata tariff plan. As handset purchases have different dynamics in ddovoloped and developing markets given carier subsidies allocated towards handsets, wo have looked at the following definitons ow pcs tare snalyses = Factory ASP of the device. This is simply the price from the mobile device vendor to either the retail or carier channel. Clearly a lower price here can have the impact of directly stimulating a lower upfront price for the consumer especially ifthe consumer ‘wants to purchase the device without a contract. '§ Subsidized ASP of the device. This concept is primarily applicable in Western Europe, North America, Latin America, and in a few markets in Asia. Mobile carrers buy the ‘ovice from the handset vendors and then offer a device subsidy in order to entice the ‘customer into a monthly contract. This makes the effective price of the device (at the point of purchase) significantly lower than the factory ASP, as shown in Figure 149 and Figure 150, This method wil stmulate demand, and especially in the case of high- fond devices, Figu 1 USS, unless otherwise sited DUS, unless eerie ‘Seuee: Company data, Creat Sues estimates ‘Seurce: Company data, Creal Susse estimates © TCO—total cost of ownership. We define this as the annualized level of total cost that 1 consumer pays in the fst year of any smariphone purchase equivalent to 12 months 149: Subsidized and Unsubsidized ASPs in NA _Figure 150: Subsidized and Unsubsidized ASPs in WE 14on1.04bn smartphone units in Tne 1.04bn amaripnone unit umber by 2015 fals reduction foramidrend 2015 based onthe BOM again fine wth ou published volume forecast fnayses—smarighane and how quick rarpnones raducon and pree band _basad on TCO and aordabity analysis 129Carprrsuss™ [ARPU (for a basic voice and data plan) plus the upfront cost of the device (if not fully subsidized or in case of no subsidies). While we believe tat the price at the point of purchase will have an impact on smartphone adoption, we increasingly note that service offerings (such as a dats service) are becoming a pivotal part of a consumer's {decision making process. For this reason, we believe that any analysis on the cost of @ smartphone must take into account the Yolal cost berne by the consumer, which, in addition to the price ofthe device, should also include the annual cost of services. Approach 1—Forecasting Smartphones on TCO [As discussed above, we fundamentally beliove that any forecast ofthe smartphone market needs to have at iis core the affordabilly and total cost of ownership (TCO) of smartphones. Given subsides in developed markets, handsets are primarily bundled with fa service package which tends to mask the true cost ofthe device. While we acknowledge nat aggressive promotions can have @ significant impact on the volume ramp-up of @ given product, we stil believe that tne TCO remains crucial for the penetration of the overall market, To arive at our TCO estimates, we have made several Important assumptions, as are detailed in Figu | Smartphone ASPs decine by some $150 by 2015. In Figure 155, we show that smartphone bill-of-materials can drop by around $85 over the period 2009-2015 (from $165 in 2009 to $82 in 2015) as chipset efficiency improves and component pricing declines, This will have drive down the ASP of a mid-end smariohone from around $250 In 2009 to $112 by 2015. "© Normal levels of ARPU decline, We have assumed annual ARPU pressure of 6% in all ‘markets (globally) to take into account pricing prassure at mobile operators. = Smartphone afordabilty threshold in each region. Based on Figure 156 and Figure 157, we arrive at smartphone affordability threshold in each region by looking at mobile telecoms as a percentage of GDP in each of the regions. As such, we have used a smartphone affordability threshold level of 1.35% in Wester Europe, 1.3% in North America, 2.2% in Centraliastern Europe and Asia-Pacific, 2.35% in Latin ‘America and 3.8% in Attica, Based on these core assumptions, we conclude that nearly 2bn people will be able to afford a smartphone by 2018, mare than doubling from our estimated addressable base of ‘885mn people in 2010. We discuss each of our assumptions in detail in the following 130131 s2yeuyso essing 19010 ‘ep Aueeung 22m0S “uwgge J aseq uoyendod ‘gessarppe nour ou Suypron uaa 20W'g40¢ Aq suoyaueus e proyeo} aige /9q IIIM ajdoad uqz punosy (¢ soununao song do} 244 “91.08 ‘A sjoys0u Bupjanep ui ugg, sajoue pue sjaeu pedojarap U auoydeus pioye oj 229 i ojdoad wugoe evonppe ue la Ly uoyanpar eruue 939 (@ 8108 0108 yeu ay eno ated yz 10 628 fa euzep avoudhws 840} 001. 28849 944 (| tore a TOL FAO, SES BORO | sires non=tr3} rome re 003 cousins our esse espe ue e268 venoves m5 esotcput pu eu, TOTS PRT OL creprrsusst™ TT ORESSTPDE PUR SST ORBIT TS OD TTL (001, Bune Aq uantig $102 Aq uaz 04 uwwggs Punosy Wo1} MosD UeD SoUOYdLELUS Jo} JOYIEW aIqeSSelPPY—SIShleUY DJ BudYA EWS :15} 21NDcarprrsuiss™ CO jn the first year is $445 on average. We analyzed the TCO for a low-end smartphone by vendor and for 19 counties as listed In Figure 153, This demonstrates that the ‘ownership cost for smariphones falls in a wide range. We believe this is owing to a number of factors such as the stage of the smartphone market in terms of nascence, Competitive dynamics, and device positioning by vendor. We also note that our analysis is based on the TCO at the lower end of the market, as we belive this subsegment of the overall market will be key to driving adoption long term. In particular, we have looked at the lowest consistent level of cost faced by a consumer by isolating comparable terminals and minimum levels of data usage. We note that the TCO vares significantly in subsidized versus unsubsidized markets, with the cost ranging from $179 to $720 with an average of around $450, Extrapolating these TCO results to other countries and regions and accounting for similarty of carierivendor dynamics, we estimate that the average global TCO is currently around $445. ‘Subsidies result in diferent components of TCO. In heavily subsidized markets like North ‘America and Western Europe, as sean in Figure 152, there is very litte upfront cost at the low-end of the smartphone market. In developing regions, however, the upfront cost ofthe device can represent a significant proportion ofthe TCO. Figure 152: TCO for a Low-End Smartphone—TCO Composition Varies Significantly by Region USS. unless otenvsesatea naa ‘Upfront Device ASP (US$) = One year ARPU (USS) o 100 200 300 400 500 600 00 Low-end smartphone TCO in 2010 (USS) ‘Seuco: Company data, Creat Suisse research Forecasting the TCO long term. As shown in Figure 152, the two components of the ‘smartphone TCO ave: |) the upfront cost of the terminal (device) to the consumer, and i) the annualized ARPU (voice and data) bundled with the device. As can be sean in subsidized markets (such as Western Europe and North America), this can be significantly different fom unsubsidized markets such as Asia Paci tesa 192creprrsusst™ s2j0unse essing ype19 Yseasal essing YPe!D ‘Hep Auedued ‘ames GOP$ 1 Wed PLE OEPS:1P BPPN Aq pamo|io} "(sauoud poipuy pus-No} aluos Jo Youre 14808) LAMB) OES Ye AIYSseUMO 1 souoydyews paseq propuy ‘syorHeU Budojanap UZ quauiBes evoydyews ayy Jo pus-mo} ul yywo16 Burp 10} fay 2 um au04 euyzep 2g pue dysieumo (098840 OL e Gurney addy ug 403809 sty ara¥eq aM "G1 S'S! YEW avoUyeUS (0698 78 diysiauno yo 1500 84614 9G] & sey abiexaAe Ue UO eP/ON BIYM GOS ‘94 0 pUuB Mo} a4 Je diysieUNO yo 4500 ety APECEID) “. — ee 7a oma TT La Ne et issn) 0s brows Bepuey —_vanendo. cuneg HS BTR PRB 1 HH SB TRIOS OFT BH OL PN HET SIO GOED TOE PIBY PE TETRIOT FST ROTTS PHOT, ‘SoujuNoD quasYIC UI ploipuy Pue “aiddy ‘eHON ‘Wik WOH} BulIaYO sUoUdyeWIS pua-MOT e 40) YsIOUM 40 09 IEIOL 2261 o3NDI 133carorrsuiss Device price willbe driven by BOM reductions. We have illustrated our assumptions on the billofmaterials decline and the ASP reductions in Figure 755. While we expect the smarghone market to evolve with new features, we believe that at the low end and rmideng, itis informative to look at how quickly the BOM may decline. Here we have made ‘assumptions that the like-for-like component BOM can decline by around 10% per annum. Allowing for other fixed costs in the total manufacturing expense and assuming some ‘gross margin pressure at the manufacturer level, we estimate the total BOM can shrink to as low at $82 by 2015. For these reasons, wa estimate thal the ASP of a smartphone can be as low as $112 by 2018, a $180 dectine from normalized 2009 levels. This allows us to then forecast the decline of the device ASP component (in the TCO calculation) by region. \We note that in subsidized markets where the upfront cost is already quite low, there wil, be minimal reductions. Expecting gradual ARPU declines. Forecasting ARPU declines is somewhat more challenging given that each carrier tends to bundle data in diferent ways with a range of voice minutes, ex's, and data. We note however, that at both the manufacturer and carrier levels, there are several intiaives to bring down costs to the consumer over time. We assume a moderate decine in the ARPU in the low-end smartphone market by 2015 based on an annualized ARPU reduction of 8% in all markets globally, Longer term TCO reduction o $234 by 2016. As shown globally, by taking into account the above assumptions for the lowend smartphone TCO (by region), we arrive at the Conclusion that the smartphone addressable market could reach close to 2bn units by 2015 in torms of affordabllty. In essence, we assume that at the very least, the reduction in the BOM will pass straight on to the consumer and this implies that the TCO for a low-end smariphone offering could reach as low as $234 longer term, a $210 decline or 47% drop trom the current TCO of $443, which implies 2 CAGR decline of 12%. Geographically, we continue fo expect a variance in TCO, even in the year 2016 as shown, In Figure 184. at “Source: Gret Suisse estimates 134_sejeuys9 essing ypoi0 wabeveg e0m0S Bie 099 gig wo. iv Wog euoydeus 'SISOOBNS s}5 1509 BsemnYOS pasyeuvou sur }oedxe OM (p spay aes (0 anp o}0z Wo 53809 ul ‘aulpep %§ 2 1520810) 394 (6 yee auo flue (conspare) eo vowpduico Buseaiou: 0 amp Kt fe sivouodis Burt, ‘Sif o S800 esteuLou Sle SS ad ‘oe “aun soe ‘ soc utenoy Be aulpap | ‘ayy jo WOK ayy oace y\(L ‘SUT EETS BOOP TY SAS PEASY WOW U9 BIOPSY BOD) $1024a z11$ yoesy pinoo suoudeWs puOpIWY © Jo ds¥ OM SIs96ENs oINsso1g WO PUE 3609 INOUOdWIOD UI Ou]D0q E UO pOsER sISAIEUY INO 59) 2:nB14 135carprrsuiss™ semaren 208 Linking our TCO Analysis to the Addressable Market [As we demonstrated above and summarized in Figure 151, the TCO for a smartphone Could realistically fall nearly 50% from the current level of $243, down to $234 by 2015; [ASP and ARPU pressure will drive this trend. To be thorough, ws will now analyze effects due to income distibution and population by region. How much are consumers prepared to spend on smartphones? Once we have estimated the cost of a smartphone to the consumer, the next issue isto calculate a sensible income threshold level. After all, while a smartphone is a desirable item, there is a limit to the income level which the average consumer would be willing to spend on the combined hardware and services. Here we have approached this analysis by looking at two separate metrics Figure 156: Mobile Telecoms as % GDP in 2009, Figuro 157: Mobile Telacoms as % GDP in 2015E ‘Sewee: Garmer, Cre Susse estimates “Source: Creat Suess estimates © Mobile spending to GDP ratio of around 1.7% globally. We see in Figure 158 that ‘globally, the amount of mobile spend varies by region, with consumers spending on average some 1.7% of GDP on mobile telephony and services. In fact, even on the assumption of a more penetrated market, as is the case in 2015, we believe that this stil wil be some 1.8% as shown in Figure 157. Hence, we would argue that the TCO ‘of a smartphone should net represent more than this level of income. In othar words, if we believe as shown in Figure 151, that the TCO of a smartphone will be $234 globally on a longer term basis, then’a user would need a minimum income level of $13,000 to afford a smartphone. This provides our key income cutoff leval fr diferent Fegions and can be used lo determine the addressable market for smartphones, In emerging markets, the lovel is highor at 2.6%. Given the 1.7% (global) and 1.4% (developed markets) threshold levels, we believe the higher 2.6% ratio for developing markets shows the importance of data and communication for users in these markets; they are prepared to spend a higher percentage of GDP on mobile telephony. In fact, {a5 shown in Figure 156, we beliave that subscribers in some emerging markets lke ‘Aitica are spending 3.8% of GDP on mobile services. Income distribution data suggests addressable market of 2bn langer term. Once we have concluded that users are prepared to spend as much as 1.3% to 3.8% of their income on smartphone purchases (depending on the geographic region), we can apply these cut off levels to determine the minimum level of GOP per head needed to purchase a smartphone based on the TCO for a low-end smartphone in that markat. Over the longer term, we then se this estimate of required income to compute the level of population that is addressable. [As shown in Figure 151, we beliove thatthe addressable market for smartphones is nearly 2bn globally, tesa 136Carprrsuss™ Smartphone Volumes to Grow at LT CAGR of 28.5% With a 2015 addressable market of some 2bn, we now turn our analysis to quantifying the rmarkst opportunity in terms of units. We conclude based upon our affordability, ponetration, and replacement analysis, tat the smartphone subscriber base longer term could be as high as 1.92bn by 2018 (around 98% penetration of the 2bn addressable market) and that the smartphone segment is set for a period of reaccelerating growth reaching volumes of 1.04bn by 2015 as shown in our summary smariphone model in igure 160. Penetration curves and the addressable market. We estimate that smariphone penetration was 24% given the installed base of 461mn smartphone subscribers at the end of 2010. The next step is to determine the evolution af this penetration on a longer term basis. Here we believe a sensible approach is to take a look at how the mobile voice market evolved. [As shown in F gure 158, the voice market followed a typical ’S" curve for penetration, as it increased from 28% to 70% over a five-year period, an increase of 40pp. Owing to the faling cost of ownership, cartier push, and consumer pull, we believe that penetration gains will be higher in smariphones than mobile voice. We estimate that smartphone effective penetration will increase from 24% in 2010 to 98% by 2015, This level of penetration would imply smartphone subscribers to the tune of 1.92bn by 2015, This likely Is conservative, since it represents a smartphone penetration reaching only i) 27% of the global mobile subscriber base and population and i) 49% of overall handsets by 2015, Figure 158: Mobile Voice Saw Penetration Gains of 40pp over Five Years After Reaching 30% jdual region reached nea 10% Mobile voice penetration ‘lobaly grew trom 2% in Yyoar 5 to 70% in yoar 10, |240pp absolute gain S yoars. ‘Sours: Company data, Creat Suisse estimates Replacement rates, Since smartphones wil operate at the higher-end of the market (versus the overall mobile industry), we assume that replacement rates in smariphones to bbe on an average 45% over 2011-2018, which remains above the global replacement rate of 27% we forecast for the overall handset industry. Looking at our regional replacement, forecasts as demonstrated in Figure 159, we expect smartphone replacement levels in North America to average around 70% in the long term. 137a corr suiss™ Figure 159: Smartphone Replacement Rate by Region epiacorent rat mensured a8 8 ofgrevious your's emariphone subscriber base Replacementrates 7008 20092010 2011 —2012e 20136 20148 70156 a ‘0% eR Tim 70M 70% 60% Go” aB% Lata ee ee ee ee es, pene MONK OHHH OOH Japan 44% BTR THOTT ATT OTH McA 42% 59% «OKT SHH HH TEN ional aTh ah “Source: Garner, Company dele, Groat Suave estimates 138se morn 2014 creprrsusst™ seyeuyso essing 4019 ‘tiep hvedwog Young ‘omnes ‘onjensesuco ofoid ys pynoo yaw uous a4} Jo} syauys9 auijor ut} Buono yey} akaOq 2M "(01 YO/0120/0120/01 10) sisyenb juaDa, Uy Sainjon auc U yprOIB Fok %98/"%96/%409/ Eb YEN "MOAB BUOAS Uo MAIA Ino spoddns i200) Ul YMOUB OK Buoas (y “9106 A syun gp Bunya ‘uy Buoy aig SG6'RG JO HYD # MOB 6] seuunjon yn ‘ABojouyy spay ses eARDEIYe SOLU Ay Jo ALO UAL! } SoUOYALELUS Daeg 2pA "IMOUB BUUN|ON Ne Wie}-F9BUO) (E “amnssoud gSy 901 Jo ads Ul ok sje dn sanuarar ut Bugsy (Kok 949 dh) wu gp Jo sjuawidys jogos {Gu 4 102 ut oI6 caunyo Buoys 998 0 anuuoD O SeO4 AUS oda aM yNqep Oye BuseoIoul Aq VON "LLOR Ul YMOIB BUUrYOA BUOAS (Z "3102 fa 9486 Buryaeot‘uowyenouiod (9008) ayqou Aq usd sem se auing-¢ fear oly ue) Jeyse} oq jum Logdope ounydys yeU9 oxDjoq OM ‘OY. Gul} fe YONI ‘su6 uoqesjaued uo pase "%r¢ 1® Mo IRS S| oTERaUad aA_ZeALA ain axaqaq aM ‘ig PUNOIE Jo YSeD—I0, Jaye KTESSAIRPE BUCH AIELIS UNe-HUy Po pale OL.0g 40 puis expe Siequosqns auoidyes RL. gp JO 9509 PAyEISU! UP Jo SIELNSS INO USN) YHE IE Mo IS UOREOUEd auOYdUEWIS (1 we ee sreluovo zroruovo Seue Sre Sim’ Shoe” yor veow ua} 1960] YOVO % $'8z I MOUB 0} JsHieUI BUdYAHEWS ‘aes SSI SIN SIO $402 Aq siun wome 4 42n0 Yoe=y 0 suawidlys ownjon auoYd}eWis jenuuy y2edx3 aM WHA] 106U07 09) 217514 139carorrsuss™ tomva sor Approach 2—Cannibalization of Other CE Markets \While many consumers will inherently use multiple devices, each fora distinct function, we believe that in aggregate, smartphones will seek to replace and effectively cannbalize 8 ‘number of other consumer electronics segments. In this section, we look af the segments that have the potential of being replaced by smariphones over time, and what the ramifications are for smartphone volumes in the long term. Portable media players. The portable media player market is dominated by the iPod, however it 's hard to see this standalone segment not being impacted especially with the sheer range of music stores now available for access via smartphones (especially with DRM-tree music). We assume that 100% of MP3 players can be cannibalzed in an ideal Personal navigation devices. Here we assume that 50% of PNDs can be potentially cannibalzed by smartphones in the long term; the quality of GPS and mapping technologies on the latest smartphones is of comparable quality Digital stil cameras. We assume that a smartphone with a robust enough camera (both in terms of pixel quality and memory capacity) wil be able to successfully replace significant percentage of point-and-shost aigtal sill cameras over time. In our analysis, we assume that all DSCs with camera qually less than 15MP can be potentially cannibalized by smartphones; tis wil be atleast 30% of the overall DSC volumes, Portable gaming devices. As shown in Figure 161, 32mn portable gaming devices are expected to be sold in 2070. However, with the onset of applications stores, and a particular focus on leveraging smartphones as gaming platforms, we believe much of the ‘market can be cannibalized Midond phones. We assume that 75% of mid-end phones selling at a price point of {$100-200 could be cannibalized by smartphones by 2015, Figure 16%: Based on Cannibalization of Adjacent Markets, Estimate Smartphone Units Could Reach 988mn by 2015 Inmilons,unoes sthoraize sates, ‘urcanbalsaton ans suggests that he smariphore mackt cous ‘Sewee: Garner IDG, Company data, Great Susse estimates In aggregate, i we fold in cannibalization of the combined revenue potential from each of these segments, we estimate that smartphone units could reach as high as 988mn by 2015, Our current market estimate of 1.04bn smariphone units (for 2015) is only around 5% higher than what is suggested by the cannibelizaion analysis; our long-term smartphone volume estimates appear reasonable in comparison, Approach 3—Price Band Work Suggests 1.04B Units A different way of looking at the long-term oppertunty for smartphones is to evaluate the price segments that these devices will realistically address over the next several years. In Figure 162, we provide @ summary ofthe overall mobile handset market by price point and note that 76% of volumes currently sell ata price under $200. Clearly the focus becomes. hhow quickly smartphones can address lower price tiers. In this analysis, we conclude that the market opportunity could be as large as 1.04bn units in the long term, given several factors tesa 140carprrsuiss™ semaren 208 Figure 162: Overall Handset Market Volumes Broken Down by Price Bands ‘Overall market spt ino $50 gre bands 2010 mark <= $50 _ $50-$100 $100-$150 $160.$700 $200-250$250-$300 $500 $950 69608400 $400-$450 5450-8500» $500 Toaluntseay) Sas ‘woftotalmarkot 3% 16% tek _— 4K HHH “Souree: Company daa, Cred Suisse estates [As shown in Figure 155, our proprietary BOM analysis suggests that the ASP of a midend ‘smartphone will kely be $112 by 2015 compared with $300 in 2008. This is a dectine of {$188 from 2009 levels, which implies around $160 from 2010 levels. As shown in Figure 463, this means that such a device could drop some 3 price bands from current levels thereby addressing the $0-50 smartphone markat by 2015 vs. the $180-200 smartphone ‘market currently. In addition, we algo assume that all mobile phones selling al an ASP of over $250 by 2015 would be smartphone. Shiting down the price bands for smartphones as shown below would suggest thatthe global smeriphone market would be 4.04bn units by 2016. Interestingly, ths also shows that nearly 60% of smartphone Industry volumes could be seling at less than $200 by 2015, Further perspective on our published estimate of 1.04bn smartphone units in 2015. This pricing Work serves to make the important point that as the smartphone segment scales to ower orces, the volume opportunity wll accelerate thereby supporting our view that global ‘smartphone volumes on an annual basis could reach our estimate of 1.04bn units by 2015, Figure 163: Our Price Band Analysis Suggests that Over tbn Smartphone Units Is Possible Over the Long Term ‘nmilons, unless othoraize sates 1) Out ata (OM) ma howe ta ASPetoremansaras can cre by anus a 1602) Oras shove ataraehae ones in rch bybovs ater nga Shahin tor fy tb, monocot fe iowa > fe re deo, smtp wlleconelon pm pm ower pce pits > sso nine wth sr > Desi of 150in marge AP mle fal snare ceca op Be pce ms, bleed ananpene omca (erontgn a mugs curry ergs pasar dSSS0 8500 cn depose > eau snargrare pratt eet prs S180. 8750 wba ot array 0m tes > Wesletelne hata prone sng lr ASP ot or tan $250 by 215 ou be ‘dle tod atan ASF fi a $0 Shetare ‘Sewee: Company data, Caner, NPO, Cre Suse estimates tesa 1aCarprrsuss™ Scaling Smartphones Based upon our extensive TCO analysis (which we have also cross-checked with our cannibalization and BOM analyses), we conclude that industry wide smartohone volumes are set to grow at a CAGR of 28.5% over the next five years (2010-2015). Furthermore, Wwe beliave that conclusions for all vendors in the handset industry, both tradiional and nontraditional, are significant. In particular, we believe that not only will smartphones ‘occupy a disproportionate share of industy value, but even newer entrants will need to ‘master the art of eperating at structurally lower price points. Rising to 49%/85% of Industry Volume/Value by 2015 Clearly our optimistic view of the smartphone market has repercussions on the wider ‘mobile handset industry. On our new estimates, we believe that longer term smartphones will become 49% of industry volume and 85% of industy value for the overall handset, rmarkst by 2015, With such an increasing and dominant share of the industry revenues coming from smartphones, we believe execution in this segment remains important and will Become crucial for the longevly of traditional handset vendors. In addition, we would highlight that strong smartphone growth does not Imply that all segments within the handset market will grow. Having updated our industry wide segmentation analysis, we artve at three main conclusions: Figure 164: Smartphones to Continue to Rise to Over 80% of Industry Value by 2015 Up from 40% in 2010, si aoe a000 070 OHNE __201RE__20TSE_a014E __201SE__CAGRIOASE Basie Phones “or «as SSB SASS To Ba% Enhanced Prones Tse 76771706 GG TMG 165% Smarphones 1997220740881 20.5% otal handset ons (mn) 7308137888 —A8TS STS —,888 0592.78 59% asc Phones Ce ee nw ahanced Phones sek 8% 9H 2H SSBB nw Smarphones 11% 19% tom 29K saa nw. Total volume mix Ga) To0% 100% 00% 100% 400% 100% 100% —_T00% ry Enhanced Prones wee 80 Smartphones soe sss aes as t98 Te Blondes ASP a a D Basie Phones 25675 24.790 26,196 26808 24974 29022 21.228 19.400 50% Enhanced Pronee s1og02 12857 102755 acct See4@ aBea5 24457 12.4 347% Total handset revenue (Son) 181,356 206,669 223,373 250,557 235,306 226,06 221,191 212.842 10% Basie Phones ) ry Emanced Prones srk 61% OK HHH ITH nw Smarphones| 25k 27% aah nw otal revenue mix) To0% 100% 100% 100% 100% 100% 100% 100% NM ‘Source: Company deta, Garner, Creat Suisse estimates Lowend to rise moderately in volume and valve terms. In Figure “84, we show that the low-end of the mobile phone market represented 32% of industry volume and 12% of industy value in 2010, Longer term, we believe that this will rise to 36% of volumes in 2018. wing to pricing pressure, however, wa expect the sogmant wil sil represent only 9% of industry revenues. We also believe that there could be an incremental 1.5bn subscribers joining mobile networks for the first time, which supports our argument for 142carprrsuiss™ low-end volume growth at the global level. This is without taking into account the replacement market in emerging markets, which we expect should grow from currently depressed levels. This, we believe, will sustain a rising unit mix for the low-end segment. ‘Smariphones to occupy the dominant share of industry value. Given our estimate of longer term smartphone market at 1.04bn units, and even allowing for price reductions. overtime, we arrive atthe conclusion that smartphones could account for 49% of industry volume and 85% of industry value by 2018. Looking at our estimate of smartphone ASPs. faling from $315 in 2010 to $174 by 2046 in Figure 164, we believe Its clear tat it remains essential fr traditeral mobile phone vendors to master the core competencies in, ‘smartphones to even survive in the market long term. ‘Mid-ond continues to shrink. Long term, we expect the midend segment to see significant declines as it falls to around 816mn units resulting in a rapidly falling industry share. Infact by 2015 we believe this segment to represent only 15%/6% of industry volumelvalue ‘compared to 4994/48% respectively in 2070. Emerging Market Increasingly Important {As the smartphone market continues to grow, we believe the segment will penetrate new regions in terms of volume. Here we would make some important observations around the mix shit geographical Figure 165: EMs Could Account for Nearly 85% of Global Smartphone Market by 2015 {2% unless thorwe states I: I | Bem 2 ‘Source: Company dala, Gartner, Groat Suisse estisios Developing regions to take a larger part of the pie. As shown in Figure °65, we can infer that some 80% of smartphone volumes came from developed markets like WE, NA and Japan in 2010, whereas developing markets accounted for only 40%. However, moving forward we see that smariphone shipments in developing regions could be as high as ‘547mn up from 121mn in 2010, which would represent §3% of the global smartphone market, We expect this to be parly driven by increasing affordability and partly by rollout of 3G mobile networks. More emphasis beyond carrier distribution. Given that litle carrier subsidization takes place in markets outside of North America, Westem Europe, and Latin America, we increasingly believe vendors will need to ether build-out their own disthibution network or sign agreements with cariersidistibutors which presents another challenge for new 143carprrsuiss™ semaren 208 tenants in the industy. We believe this holds true especially for China and India, which are among the largest handset markets globally, Low-End Rising in Importance Current we see that smartphones are predominantly @ high-end mobile device with an average ASP of $315 for 2010. In particular, looking at the price distibution of ‘smartphones in Figure 168, we see that nearly 60% of smartphone volume is priced in the range of $300 and above. In fact, as we show in an earlier section tiled “Approach 3: Price band work suggests 1.04bn units' that based upon our BOM extrapolation analysis of a typical lower end smartphone, we believe that a smartphone can be available at a price point of as low as S112 by 2016 as compared with an ASP of $300 in 2009. Based upon this analysis, we arrive atthe following important conclusions for the smartphone segment. Figure 106: Our BOM and Pricing Analyses Sugges in milons,unoes othrwisa sates ~50% of Smartphone Units Could Sell at <$200 in the Long Term 1) Out aaa (2 maj hows ha ASPs orators can dc by ach a 1602) Ou we sweats han woh ca rch byaoterereh nae Signa tn by 210, ne masala he owe: > As ce desins, erator wlacovels epee and pele overpe poh > Ou erin fn neni cars > Desde $150\n smarts ASP nla tl snatora pcg can op He pce me, ‘cops smuare cer he bs panbardgi04800 cw depo spice > vetosedsargore goers the bes povvt ater S180 8150 este eh Se > Wesle tae that a pres sg at ASP of re an $80 8 2075 wou Setore “Source: Company dat, NPD, Gro Suse osha: ‘Shiting down price ters suggests @ significant mix shift in the market. in essence, our BOM extrapolation analysis implies that each price tier could afford to shift down three price tiers as shown in Figure 165. In applying this analysis, we observe that smartphones increasingly penetrate price bands as low as $50. While we assume a healthy high-end market, nearly 50% of smartphone volumes could sell at an ASP of less than $200 in the long term. Increasing pricing pressure ahead, We believe mix shift and increased competition wil have a direct impact on pricing pressure, and as such, we assume @ pricing decline of 11% over 2010-2015. However, given our expectation for strong volume growth (CAGR of 28.5% over 2010-2015), we believe revenues with this segment can sill grow at @ CAGR: (of 14%, In addition, we would note that this pricing pressure has already begun to play out, as new players lke those listed in Figure 187 have entered the smartphone market over the past 12-18 months. Exising payers Nok, RIM, Apple, HTC, Motorola, Sarsung, LG, SonyEricsson, Pam Now players Tosh, Del Ais, Garmin, Huswe, ZTE, Ae ‘Sours: Creat Sues research tesa 144Carprrsuss™ Move Toward Lower Price Points Already on Its Way (One clear theme emerging from industry trends over the past 12-18 months is that smartphone portfolios forall vendors are rapialy evolving, and in particular, we have seen widespread industry supporto drive the market towards lower price point. Figueo 168: Specifications for Low-End Smartphone Portfolio from Different Vendors ‘Overallsmariphone portiolio _Low-end smartphone portfolio only > Vendor Bevis ‘Ava ASP (S) devices Yelow-ond Avg ASP (8) Lowost ASP (9) Wiest ASP (®) Samsung 6s 385 15 eH 183 140 240 is 4 308 6 20% 200 152 240 seme 3 395 2 5% 200 200 200 Motorie 2 282 ° 20% ar 180 240 Htc 2 442 5 19% 20 178 240 rv 1“ 20 4 20% z20 200 240 Ze 2 160 2 100% 160 10 240 deer 19 a 7 ar 200 1 224 oie 2 540 ° o% - . Totar a 3a 7 30% ee 77 BT, Towand smarighone portfolio only >> Vendor cow, 'UMTSIGSM _Touchacreon Screen size fin) Ca Wir Android Nowe 20% or 70% 28 “0%, O% Samsung 40% 0% are 28 4m% 20% sene o% 50% 00% a5 100% 100% Motorola 1% 44% 20% a4 e0% 29% Hc. on 100% 100% 29 0% 0% iva 25% s0% o% 23 100% o% ze 25% ore 92% an 25% 6% deer 18% 0% 100% 30 5% sr oe : - - . Pam : : - : Tear Ea ee rm ao 32 ae Er * The databace cud some low-and COMA and TD-SCDMA devices aa wall Source: Company data, Coat Suisse research ‘Significant low-ond launches. We see that since the beginning of 2010, key smartphone vendors have either announced or already launched a total of over 80 low-end platforms (which we define as an ASP <$250), representing 30% of these particular OEMs’ total ‘smartphone portfolios on average. In terms of core functionality for these devices, we present the key specifications in the above Figure 15%, and would highlight these particular features. © 3G has a broad representation. Of the 82 low-end smartphones launched or currently selling in the market, we estimate that 68% of devices are WCDMA, 22% are GSMIEDGE, and the remaining 10% based on other technologies (like CDMA or TD-SCDMA), "= Touchscreen seems to be the predominant form factor. Two years ago, touchscreen ‘was considered a high-end feature for smartphones. Today, however, 85% of low-end evices offer this capabilly. Interestingly, new-entrants like Huawei and Acer are 145carorrsuss™ offering touchscreen on all new low-end devices, whereas the incumbent Nokia only offers ths feature on 7 out of its 10 low-end smartphones. "= Which supports the need for larger sereon size. Even though a larger screen would materially affect the cost of bulding a lower-end device, we would point out that new lower-end devices are averaging & screen size of around 3.0". As mentioned above, ew envants like Huawei, ZTE, and Acer offer the largest average screen sizes (at 3.2'/3.1°13.0" respectively) in the low-end "= Not all features are emphasized in the low-end as yet. AS shown in Figure 188, features like WiFi, which are generally standard in higher-end smartphones, stil only ‘account for 61% of low end smariphone portfolios, = Factory ASP of $189 and probably going lower. Already we see that low-end smartphones inodueed over the last 12 months have an average factory ASP of $189, with Huawoi and ZTE already announcing Android based terminals at less than $100. The wireless industry is supporting the move toward lower end. While device launches have clearly demonstrated the trend toward lower price points, we equally acknowledge nat across the wireless supply chain ranging from vendors (lke Nokia, HTC. and Samsung) to wireless operators (euch as Orange and China Unicom) and chipset Companies (lke MediaTek ~ covered by Credit Suisse Asian Foundries analyst Randy ‘Abrams, Marvell, and Broadcom - covered by Credit Suisse Semiconductor analyst John Pitzer, i clear thatthe industry has agreed for the need fo grow this lower-end segment fof the market, We have highlighted these recent comments in Figure 169) 146carorrsuss™ Figure 169: We See Increasing Signs of Low-End Smartphones Being Embraced Across the Wireless Industry Date Company Comments ‘-Feb-10 Broadcom Designed n Brondconvs proven 65 narcmetar GMOS proces, te naw BOMBTSGS sales mandactrera To bala low cost, lw power, next generation 3G HSUPA phones wih breaktroughfealures, sleek frm factors and very long baer wes DS-FGD-TO WediaTak —Morosoh ane MediaTek announce parhorship fo provide a mulbwedia ih sma phone solionpriarly aimed st EM Ae rsult ofthis dal, consumers wl beable fo buy a wide range ofetuteren Windows phonas wit the beet cost performance rato inne mar TOFeE-10 STEissson “Our UST platorm has been designed to enable th smaighons io Brak au oflscuvent high-end vizhe and come ae mass-market product in 2010"= Mare Ctl, Head of 3G and Mutinedia at ST-Crexaan 12Feb-10 Marvel “Marvels daring Be plato“ for $99 amatprone that provides instant access, ive contort, Nigh performance” 30 mobile gaming, ren HO mess and a weath of appestons"- Well Da, GM of Consumer & Computing Business at Marvel TEFoG-10Intnaar Ininon announced the avalaily af XM26, the aaa platorm le SG slow modem Tamiy.by aang anced HSPA- features, wie significant redcing board space, power consumption and BOM cost, ieFeei0 Ze ZTE unveiled to low-cost Mobil louchse'sen Vary Tueh I, avaible to UK propay users om Waren. Kw fatal for £59.98 and supports EDGE technology. TE Fab-10 Grange “in 2010 we wil be oflaing an Android handset a €150 on prepay, and | would expec al w wil have an Andro” handset a below €750 before Chssimas. The was considered unachievable six mons ago": Yves Maite, SUP of Noble Mutimeda and Devioes st Orga. 17-60-10. China Unio *. we nee beter handeste fo put 36 ine fu use. We already have deviogs athe nag-and, but we now plan to launch a low-cost 36 smartphone inthe fate” La Yimin, Prosser af China Unicom . 17-Fob-10 Huawei “Key io unlocuing he poten! usage of 3G he need fo a USS "50 smartphone wih IPhane-ke capable Guo Ping, Cnatman of Huawe! Devices . 17-66-10 Symbian “The rst Symbian smarphonos wits unsubs died prices of €100 wil rach the markt his yar. year wo wil ZeReO Nokia "in i Nokia dolvored revenue and earings growth on a yoar over yoar bass. out dovcos and voncas busness, Noiacontruss to show soit smarigtone momentum nlower pce points" at-Merto RIM "your assumaton that ow ASP products, Is whats driving inematonal corinly 8520 has been success has ped e8scallystin some of hose marks. and 8520 also had sgl gross marin" - Ede! Ebbs, IR st IM Beret the smartphone segment is conning Rs sor momentum. and it aking an Wereasing share of what ypiealy has boen tne feature phove sogrret. wore also pleased fo soe more and more operators expert with {or 36 data serdoes”- Paul Jacabs, Charman and CEO st Quaicomm Boato Nowa 7 arasing products for European matkss 6 wel, bt | can cally give ai futher ASP dynamics regarding those products" = Timo inamvotia, CFO at Nowa, 26-0610 Broadcom “Inthe rest ofthe wor, people are looking for cost-ofactv smartphones, ard smartphones they ean buy ata rote afrsable pce" - Seal MeGragor,Prescet and CEO at Broadcom, reocso HTC inthe Eutopoan mate, we are ses ou mi ers produ! seeing a value, a6 nested before. S01 hk a ‘fhe racine forthe mi ers pring, | Pink we ate much mre ready han we were years ago" Hung BOK) Woiwee ‘ny ove poabledeice wid an acoiandly dosiga orlrstine smarghone users ACRTET, we launched vee mae devices ~ Bravo, FLPSIDE and FLIPOUT. With a vary of Scroon sizes and farm factrs and aforéaboretal price paints, these devices give consumer adéod chles in tit. tmarghone sxperionces"- Sanjay Jha, CEO of Mobile Devices at Motors. a3Nowt0 “We continue to see upside inthe high end of fe marke, a8 wall 8 new wave of opporuniy coming nthe frm of” lower ost hgh volume 3G phonesin bon developed and developing geogaphn, Te cos aovantages of cur Inegtatd approach and latorm competoliy among our products posions us lo gow as hese vends ay auth tn markaiplace"- Stove Mollonkopt, EVP, Present at Qualemm, O5en-1t HTC /ATAT announcod HTC Frostye alow cost BREW smaripone. Froesie specs include 32 Inch stron, S12NB Ineral storage, 528MM Quateomm processor I supports GSMIUNTSIHSPA and comes wit HTC Sense Ul orven-tt Ze ZTE launched a we variety of Anois davies i the yea of 2010 incucng the Liberehandsstn Japas, ine Ble handstn France, te Racer Manat n te U Kane the LIGHT abet in Europe and Asia. The company contnuos to dela Inovatve products that take advantage othe most advanced sofware onthe market at an Unsurpassed vali to consumers . 2ivantt HTC *S0 we wil continue to, 0 tom the partiolo perspective, s you can se and sugges fom out 4Q ASP and at tho ‘moment Qi think wo 60 see cominued quis good demand on ourhig-and aoduct ‘or the year, and atncely tre sat nspiation rom our Wie, We willensnus to work onthe mid er product ta luther penetra the whole Smarghone markt - Joey Chong, Distr of Rat HTC. Bé-JarcTt Qualcomm "We do s66 alt of movement owards mass-markt smarghones tal shoud fab arow tha maial Butea consequence als lower the average sling pice a Bil Kats, EVP & CFO at Qualcomm. Ga-Feb-11 Broadcom — Our BON21890 processor ands Broadstone” rteree pallor make teaser han aver for OFWa to brng markt ate smariphone features that people wantin an afordabie handset designs. TEFSE-TT ModiaTek "We toveve tat ournew M8573 platiorm enables our customs to most hs consumer demand and bing unprecedented level of perfomance a lwer-areed smartphones Jefley Ju, GM, Smartghane at Medak ‘Source: Company daa, Great Suisse research “a7carorrsuss™ tomva sor Lower-end chipsets should support the expansion. A clear theme in the past few months has also been from the supply chain and particularly chipset companies to support this move to lower price points. As shown in Figure ‘70, MediaTek, ST-Eriesson, Broadcom, Infineon and Marvell have all developed lower-end solutions to specificaly address this {rowing segment of the market. Figure 170: Landscape of Lower-End Chipsets Recently Announced Dato of Anne —aFen09 Bepi0 15.eb-10 0 Technology supported OSE. HSDPA HSOPAISUPA—HSDPANSUPA : 5100515 300 Sue eye z Yeuves is Video suppor : HGR ARM ear : ARM : ARMS ‘Source: Company data, Great Suisse research Tiered data pricing. As we discuss later in this repor, from an affordabilty perspective, while having lower-end smartphone price points is seen as important from 8 handset vendor's perspective, we acknowledge that the annualized ARPU level sill remains & barrie to afferdabilty In this context, we believe is reassuring that wireless carers such {as Vodafone, AT&T, and Verizon (covered by Credit Suisse Telecommunication Services analyst Jonathan Chaplin) all seem to acknowledge the need to introduce tiered data pricing. We believe that this move is being driven by the experience of carriers such as ATT where management has previously noted that 3% of users are accountng for 40% ofall aff tesa 148carprrsuiss™ Figure 171: Several Industry Vendors Suggest Trends Are Moving to Accommodate Tired Data Pricing Date Company ‘Comments Bends Vetzon “We els ecenly expanded aur dala pcg plans oWlude wo pew lowe ost, usage based oplors for customers wih feature phones wh may rl necessarily want to pay for unlimited data - John Kia, CFO at Yerson ieFeei0 Arar “Evontvaly, ere needs io be @ model hat adsrenees the United spectam lave and suppor coninaed rout in mobile sroscband usage innovalan and ivestmont- John Donovan, CTO oF ATAT, ting that he - Sperator may move sway from fata "alyou-can-2at packages. TEFeE-10 Research n Wot *-Manutactrers naa bot sla busing mo efficent epplcatons and mors oficen envces.fwe dant start conserving banewah, nthe next few years we are going to run nfo capac erneh, You are already lxponencing he eapaciy eunen inthe US" ke Lazar, co-CEO of RIM iTFee-10Ercason High speod networks a6 27 enable but we must rake them ease o manage, and move away FOR simpli fatvatechigig” -Hans Vestaeg, CEO of Eresson T7FoG-10 China Unicom “nthe nea uur Nate wil Become a common problem fr aporalar, ana wave baking to mplament atorative pricing models for mob broadband” - Lu Yimin, Present of China Unicom. irFee-10 KODI *.-prsont policy of fatale daa tari =a big Headache. However, mplometng muloieiovel dala prcha {or businass users has proven tobe very complex.."-Tadashl Onadera, Chatman of KOD. TFFeb-10 TelaSonera “Telecom operaor wil need 1 adopt vaiable pricing model for modte dat a8 network wai Vass end capaci gos stranos"- Kenneth Karborg, Head of Wabily Sercos at TelaSorara, 7-Feb-10 ‘Ales Lvcant—* tho industry must seain to move away rm he ‘al-yo-

You might also like

• Coatue April2024

  

  No ratings yet

  Coatue April2024

  68 pages
• 2023121200041

  

  No ratings yet

  2023121200041

  84 pages
• Essential Colour Guide PDF

  

  88% (8)

  Essential Colour Guide PDF

  74 pages
• 高盛：中国-"L 型"房地产行业复苏势在必行

  

  100% (1)

  高盛：中国-"L 型"房地产行业复苏势在必行

  14 pages
• Goldman Sachs AI Adoption Tracker 2024Q2 Hardware Investment Grows

  

  No ratings yet

  Goldman Sachs AI Adoption Tracker 2024Q2 Hardware Investment Grows

  13 pages
• What To Expect From Tech

  

  No ratings yet

  What To Expect From Tech

  20 pages
• Global_M_A_Report_2025_1741195451

  

  No ratings yet

  Global_M_A_Report_2025_1741195451

  88 pages
• 2023112700022

  

  No ratings yet

  2023112700022

  117 pages
• Petrochemicals: Preparing For A Supercycle

  

  No ratings yet

  Petrochemicals: Preparing For A Supercycle

  77 pages
• Kelun Biotech (6990.HK) - Early-Mover ADC Player in China With Global Potential Initiate at Buy

  

  No ratings yet

  Kelun Biotech (6990.HK) - Early-Mover ADC Player in China With Global Potential Initiate at Buy

  75 pages
• India Internet - A Closer Look Into The Future

  

  No ratings yet

  India Internet - A Closer Look Into The Future

  108 pages
• I M R C ' H D M S: Ndependent Arket Esearch Hina S OME Ining Arket Tudy

  

  No ratings yet

  I M R C ' H D M S: Ndependent Arket Esearch Hina S OME Ining Arket Tudy

  32 pages
• Global Market Views Lower Macro Tails, Higher Political Tails

  

  No ratings yet

  Global Market Views Lower Macro Tails, Higher Political Tails

  13 pages
• GS On CHF Research

  

  No ratings yet

  GS On CHF Research

  9 pages
• Software latest update

  

  No ratings yet

  Software latest update

  22 pages
• 高盛 生成式AI投资框架（英）

  

  No ratings yet

  高盛 生成式AI投资框架（英）

  61 pages
• Cyber 02

  

  No ratings yet

  Cyber 02

  72 pages
• DB_Macro_Markets_Tech_1734792159

  

  No ratings yet

  DB_Macro_Markets_Tech_1734792159

  30 pages
• Semiconductors Outlook - JPM

  

  No ratings yet

  Semiconductors Outlook - JPM

  39 pages
• Michael Guichon Sohn Conference Presentation

  

  No ratings yet

  Michael Guichon Sohn Conference Presentation

  49 pages
• Global Strategy Paper - Why AI Is Not A Bubble

  

  No ratings yet

  Global Strategy Paper - Why AI Is Not A Bubble

  37 pages
• Moody - Finance Companies Methodology

  

  No ratings yet

  Moody - Finance Companies Methodology

  48 pages
• Music in The Air - Music Still Sounds Good in A Macro Downturn Raising Global Industry Forecasts

  

  No ratings yet

  Music in The Air - Music Still Sounds Good in A Macro Downturn Raising Global Industry Forecasts

  64 pages
• Global Economics Analyst_ DeepSeek Raises Micro Risks, Macro Upside (Briggs)

  

  No ratings yet

  Global Economics Analyst_ DeepSeek Raises Micro Risks, Macro Upside (Briggs)

  11 pages
• Barclays-NVIDIA 20240814

  

  No ratings yet

  Barclays-NVIDIA 20240814

  21 pages
• Lucy Lamb, Doctor'S Wife: Sara Seale

  

  100% (1)

  Lucy Lamb, Doctor'S Wife: Sara Seale

  179 pages
• 2025 Year-Ahead Investment Outlook

  

  No ratings yet

  2025 Year-Ahead Investment Outlook

  16 pages
• ASEAN Banks 2025 outlook Navigating a shallower rate cut environment; 3 rating changes

  

  No ratings yet

  ASEAN Banks 2025 outlook Navigating a shallower rate cut environment; 3 rating changes

  59 pages
• Navigating China Internet - Lifting Preference For Advertising On Faster Re-Opening 7 Key Themes Buy Alibaba (Add To CL), Te...

  

  No ratings yet

  Navigating China Internet - Lifting Preference For Advertising On Faster Re-Opening 7 Key Themes Buy Alibaba (Add To CL), Te...

  92 pages
• Americas Technology - Internet - Q3 '22 Gaming Industry Preview - Headwinds Remain in '22 Industry Themes Likely Drive 2023 & ...

  

  No ratings yet

  Americas Technology - Internet - Q3 '22 Gaming Industry Preview - Headwinds Remain in '22 Industry Themes Likely Drive 2023 & ...

  21 pages
• Upl PDF

  

  No ratings yet

  Upl PDF

  44 pages
• Morningstar Report-636d1777e26a17cb579912df

  

  No ratings yet

  Morningstar Report-636d1777e26a17cb579912df

  22 pages
• May 18, 2023 - Investor Day Slide Presentation

  

  No ratings yet

  May 18, 2023 - Investor Day Slide Presentation

  14 pages
• Morningstar Report-636d1777e26a17cb579912df

  

  No ratings yet

  Morningstar Report-636d1777e26a17cb579912df

  21 pages
• Nomura - Oct 11, 12 - International Brands - Global Retail Databook

  

  No ratings yet

  Nomura - Oct 11, 12 - International Brands - Global Retail Databook

  84 pages
• 2011-01-03 JPM (2011 Internet Investment Guide)

  

  No ratings yet

  2011-01-03 JPM (2011 Internet Investment Guide)

  416 pages
• NIO Inc. (NIO_9866.HK)_ Relatively Light New Model Pipeline With Potential Cost Tailwinds; Initiate at Neutral(1)

  

  No ratings yet

  NIO Inc. (NIO_9866.HK)_ Relatively Light New Model Pipeline With Potential Cost Tailwinds; Initiate at Neutral(1)

  28 pages
• Ame Research Thesis 11-13-2014

  

  No ratings yet

  Ame Research Thesis 11-13-2014

  890 pages
• Predictions

  

  No ratings yet

  Predictions

  2 pages
• PT. Mirae Asset Sekuritas Indonesia - Initiation Report PDF

  

  No ratings yet

  PT. Mirae Asset Sekuritas Indonesia - Initiation Report PDF

  75 pages
• GS Bayer AG

  

  No ratings yet

  GS Bayer AG

  7 pages
• Accenture Vendor Rating

  

  No ratings yet

  Accenture Vendor Rating

  9 pages
• Deutsche Bank - Cloud Computing. Clear Skies Ahead

  

  No ratings yet

  Deutsche Bank - Cloud Computing. Clear Skies Ahead

  20 pages
• 2021-11-09-SWIR - OQ-RBC Capital Markets-Normalizing Following Manufacturing Disruptions-94496714

  

  No ratings yet

  2021-11-09-SWIR - OQ-RBC Capital Markets-Normalizing Following Manufacturing Disruptions-94496714

  14 pages
• DigiCapital Global Video Games Investment Review

  

  No ratings yet

  DigiCapital Global Video Games Investment Review

  69 pages
• 2022 - Strategy Matters - The Equity Risk Premium (ERP) in The Post-Modern Cycle - Cyclical Versus Structural Trends

  

  No ratings yet

  2022 - Strategy Matters - The Equity Risk Premium (ERP) in The Post-Modern Cycle - Cyclical Versus Structural Trends

  23 pages
• Mizuho Securities GPN SQ We Applaud Strong EBITDA COIN More of The Same GP

  

  No ratings yet

  Mizuho Securities GPN SQ We Applaud Strong EBITDA COIN More of The Same GP

  8 pages
• NCBC SaudiFactbook 2012

  

  No ratings yet

  NCBC SaudiFactbook 2012

  232 pages
• JP MORGAN Pakistan Analysis

  

  No ratings yet

  JP MORGAN Pakistan Analysis

  64 pages
• Bit Coin Mini Case Solutions - Book by Arthur I. Stonehill, David K. Eiteman, and Michael H Moffett

  

  No ratings yet

  Bit Coin Mini Case Solutions - Book by Arthur I. Stonehill, David K. Eiteman, and Michael H Moffett

  5 pages
• DB CHTR

  

  No ratings yet

  DB CHTR

  24 pages
• Morgan Stanley Europe Update

  

  No ratings yet

  Morgan Stanley Europe Update

  25 pages
• Part2-EquityInvestmentsAsAHedgeAgainst LazardResearch

  

  No ratings yet

  Part2-EquityInvestmentsAsAHedgeAgainst LazardResearch

  7 pages
• Dorman - Jefferies

  

  No ratings yet

  Dorman - Jefferies

  20 pages
• Mckinsey Appraisal - Appraisal

  

  No ratings yet

  Mckinsey Appraisal - Appraisal

  8 pages
• Computer Revision Class3

  

  50% (2)

  Computer Revision Class3

  4 pages
• HSBC - Currency Outlook December

  

  No ratings yet

  HSBC - Currency Outlook December

  46 pages
• Revision Material For Eng Lit Icse Class 10

  

  No ratings yet

  Revision Material For Eng Lit Icse Class 10

  60 pages
• VALUExVail 2012 James Chanos

  

  No ratings yet

  VALUExVail 2012 James Chanos

  30 pages
• Becton Dickinson BDX Thesis East Coast Asset MGMT

  

  100% (1)

  Becton Dickinson BDX Thesis East Coast Asset MGMT

  12 pages
• G10 Annual Outlook

  

  No ratings yet

  G10 Annual Outlook

  27 pages
• Gsma Me Apac 2014

  

  No ratings yet

  Gsma Me Apac 2014

  76 pages
• Grade9 2012

  

  No ratings yet

  Grade9 2012

  31 pages
• Quarterly Market Review: Second Quarter 2016

  

  No ratings yet

  Quarterly Market Review: Second Quarter 2016

  15 pages
• Workbook

  

  No ratings yet

  Workbook

  8 pages
• European World-Class Results Are Close To Overall Results, But Fewer Top Performers On Cost

  

  No ratings yet

  European World-Class Results Are Close To Overall Results, But Fewer Top Performers On Cost

  35 pages
• EnableSalestoElicitMktIntellMay08 PDF

  

  No ratings yet

  EnableSalestoElicitMktIntellMay08 PDF

  2 pages
• The Living Organisms

  

  No ratings yet

  The Living Organisms

  16 pages
• Enable Salesto Elicit MKT Intell May 08

  

  No ratings yet

  Enable Salesto Elicit MKT Intell May 08

  2 pages
• MannDeshi Entrepreneurs Program

  

  No ratings yet

  MannDeshi Entrepreneurs Program

  15 pages
• Joining The Mobile Payments Chase: Second City National Bank

  

  No ratings yet

  Joining The Mobile Payments Chase: Second City National Bank

  18 pages
• Net Cracker

  

  No ratings yet

  Net Cracker

  10 pages
• HHW Class IV

  

  No ratings yet

  HHW Class IV

  5 pages
• Revision Sheet Science

  

  No ratings yet

  Revision Sheet Science

  6 pages
• Opening Closing Lines1

  

  No ratings yet

  Opening Closing Lines1

  2 pages
• Delhi Public School. Sushant Lok Social Science Class-Iii Revision Worksheet - 1

  

  No ratings yet

  Delhi Public School. Sushant Lok Social Science Class-Iii Revision Worksheet - 1

  2 pages
• LEAD Program Poster Summer 2018

  

  No ratings yet

  LEAD Program Poster Summer 2018

  1 page
• Fganh: 1 Vuqpnsn Ys (Ku 2 Ikb Iqlrd Forku@Dgkuh Lap Ij VK/KKFJR Vifbr XN Ka'K

  

  No ratings yet

  Fganh: 1 Vuqpnsn Ys (Ku 2 Ikb Iqlrd Forku@Dgkuh Lap Ij VK/KKFJR Vifbr XN Ka'K

  1 page
• Communications Brief Template

  

  No ratings yet

  Communications Brief Template

  1 page
• Class V G.Science Q.1 Fill in The Blanks

  

  No ratings yet

  Class V G.Science Q.1 Fill in The Blanks

  5 pages
• Out of The Box: Elevance Implicity AND Eciprocity

  

  No ratings yet

  Out of The Box: Elevance Implicity AND Eciprocity

  2 pages
• The Lever

  

  No ratings yet

  The Lever

  2 pages
• Common Stock Valuation Irc 409a Valuation Valuation and Advisory Case Study

  

  No ratings yet

  Common Stock Valuation Irc 409a Valuation Valuation and Advisory Case Study

  3 pages