U.S.

Solar Photovoltaic System

Cost Benchmark: Q1 2017
Ran Fu, David Feldman, Robert Margolis,
Mike Woodhouse, and Kristen Ardani
National Renewable Energy Laboratory

NREL is a national laboratory of the U.S. Department of Energy

Office of Energy Efficiency & Renewable Energy
Operated by the Alliance for Sustainable Energy, LLC
This report is available at no cost from the National Renewable Energy
Laboratory (NREL) at www.nrel.gov/publications.

Technical Report
NREL/TP-6A20-68925
September 2017

Contract No. DE-AC36-08GO28308

 U.S. Solar Photovoltaic System
Cost Benchmark: Q1 2017
Ran Fu, David Feldman, Robert Margolis,
Mike Woodhouse, and Kristen Ardani
National Renewable Energy Laboratory
Prepared under Task No. SETP.10308.03.01.10

NREL is a national laboratory of the U.S. Department of Energy

Office of Energy Efficiency & Renewable Energy
Operated by the Alliance for Sustainable Energy, LLC
This report is available at no cost from the National Renewable Energy
Laboratory (NREL) at www.nrel.gov/publications.

National Renewable Energy Laboratory Technical Report

15013 Denver West Parkway NREL/TP-6A20-68925
Golden, CO 80401 August 2017
303-275-3000 • www.nrel.gov
Contract No. DE-AC36-08GO28308
 NOTICE

This report was prepared as an account of work sponsored by an agency of the United States government.
Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty,
express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of
any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation,
or favoring by the United States government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States government or any agency thereof.

This report is available at no cost from the National Renewable Energy

Laboratory (NREL) at www.nrel.gov/publications.

Available electronically at SciTech Connect http:/www.osti.gov/scitech

Available for a processing fee to U.S. Department of Energy

and its contractors, in paper, from:

U.S. Department of Energy

Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831-0062
OSTI http://www.osti.gov
Phone: 865.576.8401
Fax: 865.576.5728
Email: [email protected]

Available for sale to the public, in paper, from:

U.S. Department of Commerce

National Technical Information Service
5301 Shawnee Road
Alexandria, VA 22312
NTIS http://www.ntis.gov
Phone: 800.553.6847 or 703.605.6000
Fax: 703.605.6900
Email: [email protected]

Cover Photos by Dennis Schroeder: (left to right) NREL 26173, NREL 18302, NREL 19758, NREL 29642, NREL 19795.

NREL prints on paper that contains recycled content.

 List of Acronyms
AC alternating current
BOS balance of system
DC direct current
EPC engineering, procurement, and construction
FICA Federal Insurance Contributions Act
GW gigawatt
ILR inverter loading ratio
ITC investment tax credit
LCOE levelized cost of energy
MACRS Modified Accelerated Cost Recovery System
MLPE module-level power electronics
NEC National Electric Code
NEM net energy metering
NREL National Renewable Energy Laboratory
O&M operation and maintenance
PERC passivated emitter and rear cells
PII permitting, inspection, and interconnection
PV photovoltaic(s)
Q quarter
R&D research and development
SAM System Advisor Model
SG&A sales, general, and administrative
TPO third party ownership
USD U.S. dollars
Vdc volts direct current
Wac watts alternating current
Wdc watts direct current

iii
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Executive Summary
This report benchmarks U.S. solar photovoltaic (PV) system installed costs as of the first quarter
of 2017 (Q1 2017). We use a bottom-up methodology, accounting for all system and project-
development costs incurred during the installation to model the costs for residential, commercial,
and utility-scale systems. In general, we attempt to model the typical installation techniques and
business operations from an installed-cost perspective. Costs are represented from the
perspective of the developer/installer; thus, all hardware costs represent the price at which
components are purchased by the developer/installer, not accounting for preexisting supply
agreements or other contracts. Importantly, the benchmark also represents the sales price paid to
the installer; therefore, it includes profit in the cost of the hardware, 1 along with the profit the
installer/developer receives, as a separate cost category. However, it does not include any
additional net profit, such as a developer fee or price gross-up, which is common in the
marketplace. We adopt this approach owing to the wide variation in developer profits in all three
sectors, where project pricing is highly dependent on region and project specifics such as local
retail electricity rate structures, local rebate and incentive structures, competitive environment,
and overall project or deal structures. Finally, our benchmarks are national averages weighted
by state installed capacities. Table ES-1 summarizes the first order benchmark assumptions.

Table ES-1. Benchmark Assumptions

Unit Description

Values 2017 U.S. dollars (USD)

System Sizes In direct current (DC) terms; inverter prices are converted by DC-to-alternating
current (AC) ratios.

PV Sector Description Size Range

Residential Residential rooftop systems 3–10 kW

Commercial Commercial rooftop systems, ballasted racking 10 kW–2 MW

Utility-Scale Ground-mounted systems, fixed-tilt and one-axis tracker >2 MW

Based on our bottom-up modeling, the Q1 2017 PV cost benchmarks are:

• $2.80 per watt DC (Wdc) (or $3.22 per watt AC [Wac]) for residential systems
• $1.85/Wdc (or $2.13/Wac) for commercial systems
• $1.03/Wdc (or $1.34/Wac) for fixed-tilt utility-scale systems
• $1.11/Wdc (or $1.44/Wac) for one-axis-tracking utility-scale systems. 2

1
Profit is one of the differentiators between “cost” (aggregated expenses incurred by a developer/installer to build
a system) and “price” (what the end user pays for a system).
2
This year, we use the same DC-to-AC ratio (1.3) for both fixed-tilt and one-axis-tracking utility-scale PV systems
(see Section 2.5).

iv
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Overall, modeled PV installed costs declined, year over year, in Q1 2017 for all three sectors,
as they have done each year since we began modeling PV system costs.

Figure ES-1 puts our Q1 2017 benchmark results in context with the results of previous NREL
benchmarking analyses. When comparing the results across this period, it is important to note
the following:

1. Values are inflation adjusted using the Consumer Price Index. Thus, historical values from
our models are adjusted and presented as real USD instead of nominal USD.
2. Cost categories are aggregated for comparison purposes. “Soft Costs – Others” represents
permitting, inspection, and interconnection (PII); land acquisition; sales tax; and engineering,
procurement, and construction (EPC)/developer overhead and net profit.
3. The “Utility-Scale PV, One-Axis Tracker (100 MW)” consists of our previous bottom-up
results (2010 and 2013–2016) and interpolation estimates for 2009 and 2011–2012.
4. A comparison of Q1 2016 and Q1 2017 is presented in Table ES-2.

v
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure ES-1. NREL PV system cost benchmark summary (inflation adjusted), 2010–2017
The inflation-adjusted system cost differences between Q1 2016 and Q1 2017 are $0.18/Wdc (residential), $0.32/Wdc (commercial),
and $0.42/Wdc (fixed-tilt utility-scale). Table ES-2 shows the benchmarked values for all three sectors and drivers of cost decrease
and increase.

vi
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table ES-2. Comparison of Q1 2016 and Q1 2017 PV System Cost Benchmarks

Sector Residential PV Commercial PV Utility-Scale PV, Fixed-Tilt

Q1 2016 Benchmarks in
$2.93 $2.13 $1.42
2016 USD/Wdc

Q1 2016 Benchmarks in
$2.98 $2.17 $1.45
2017 USD/Wdc

Q1 2017 Benchmarks in
$2.80 $1.85 $1.03
2017 USD/Wdc

• Lower module • Lower module • Lower module price

price price • Lower inverter price
• Lower inverter • Lower inverter • Higher module
price price efficiency
• Higher module • Higher module
efficiency efficiency
• Lower electrical • Smaller
BOS commodity developer team
Drivers of Cost Decrease price
• Higher small
installer market
share
• Lower sales &
marketing costs
• Lower overhead
(general &
administrative)

• Higher labor • Higher labor • Higher labor wages

wages wages • Higher net profit to
• Higher advanced • Higher PII costs EPC/developer
inverter adoption • Higher net profit
Drivers of Cost Increase • More BOS to
components for EPC/developer
rapid shutdown
• Higher supply-
chain costs

vii
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 As Figure ES-1 shows, hardware costs—and module prices in particular—declined substantially
in Q1 2017 owing to an imbalance in global module supply and demand. This has increased the
importance of non-hardware, or “soft,” costs. 3 Figure ES-2 shows the growing contribution from
soft costs. 4 Soft costs and hardware costs also interact with each other. For instance, module
efficiency improvements have reduced the number of modules required to construct a system
of a given size, thus reducing hardware costs. This trend has also reduced soft costs from direct
labor and related installation overhead.

Figure ES-2. Modeled trend of soft cost as a proportion of total cost by sector, 2010–2017

Also, our bottom-up system cost models enable us to investigate regional variations, system
configurations (such as MLPE vs. non-MLPE, fixed-tilt vs. one-axis tracker, and small vs. large
system size), and business structures (such as installer vs. integrator, and EPC vs. developer).
Different scenarios result in different costs, so consistent comparisons can only be made when
cost scenarios are aligned.

Finally, the reductions in installed cost, along with improvements in operation, system design,
and technology have resulted in significant reduction in the cost of electricity, as shown in
Figure ES-3. U.S. residential and commercial PV systems are 86% and 89% toward achieving
SunShot’s 2020 electricity price targets, and U.S. utility-scale PV systems have achieved their
2020 SunShot target three years early.

3
Soft cost = total cost - hardware (module, inverter, structural and electrical BOS) cost.
4
An increasing soft cost proportion in Figure ES-2 indicates soft costs declined more slowly than did hardware
costs; it does not indicate soft costs increased on an absolute basis.

viii
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure ES-3. NREL PV LCOE benchmark summary (inflation adjusted), 2010–2017

ix
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table of Contents
1 Introduction ........................................................................................................................................... 1
2 Model Inputs and Sources ................................................................................................................... 3
2.1 California’s NEM Interconnection Applications Data Set ............................................................ 3
2.2 Module Power and Efficiency ....................................................................................................... 4
2.3 PV System Size ............................................................................................................................. 5
2.4 Module-Level Power Electronics .................................................................................................. 5
2.5 Inverter Price and DC-to-AC Ratios ........................................................................................... 10
2.6 Module Prices.............................................................................................................................. 12
2.7 Small Installers vs. National Integrators in the Residential PV Model ....................................... 16
3 Residential PV Model ......................................................................................................................... 18
3.1 Residential Model Structure, Inputs, and Assumptions .............................................................. 18
3.2 Residential Model Output ........................................................................................................... 20
3.3 Residential Model Output vs. Reported Costs ............................................................................ 22
3.4 Residential PV Price Benchmark Historical Trends ................................................................... 23
3.5 Residential PV Levelized Cost of Energy Historical Trends ...................................................... 24
4 Commercial PV Model ........................................................................................................................ 27
4.1 Commercial Model Structure, Inputs, and Assumptions............................................................. 27
4.2 Commercial Model Output.......................................................................................................... 29
4.3 Commercial PV Price Benchmark Historical Trends.................................................................. 31
4.4 Commercial PV Levelized Cost of Energy Historical Trends .................................................... 31
5 Utility-Scale PV Model ........................................................................................................................ 34
5.1 Utility-Scale Model Structure, Inputs, and Assumptions ............................................................ 34
5.2 Utility-Scale Model Output ......................................................................................................... 37
5.3 Utility-Scale PV Price Benchmark Historical Trends ................................................................. 40
5.4 Utility-Scale PV Levelized Cost of Energy Historical Trends .................................................... 41
6 Model Applications ............................................................................................................................. 44
6.1 System Cost Reduction from Economies of Scale ...................................................................... 44
6.2 Module Efficiency Impacts ......................................................................................................... 44
6.3 Regional LCOE ........................................................................................................................... 46
7 Conclusions ........................................................................................................................................ 48
References ................................................................................................................................................. 52
Appendix A. Historical PV System Benchmarks in 2010 USD ............................................................. 54
Appendix B. PV System LCOE Benchmarks in 2017 and 2010 USD ................................................... 57

x
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 List of Figures
Figure ES-1. NREL PV system cost benchmark summary (inflation adjusted), 2010–2017 ...................... vi
Figure ES-2. Modeled trend of soft cost as a proportion of total cost by sector, 2010–2017 .................... viii
Figure ES-3. NREL PV LCOE benchmark summary (inflation adjusted), 2010–2017 .............................. ix
Figure 1. U.S. PV market growth, 2004–2016, in gigawatts of direct-current (DC) capacity
(Bloomberg 2017) .................................................................................................................... 1
Figure 2. Installed capacities of residential and commercial PV systems covered by the California
NEM database (Go Solar CA 2017) compared with GTM data (GTM Research 2017),
2010–2016................................................................................................................................ 4
Figure 3. Module power and efficiency trends from the California NEM database (Go Solar CA 2017),
2010–2016................................................................................................................................ 4
Figure 4. PV system size trends from the California NEM database (Go Solar CA 2017), 2010–2016 ...... 5
Figure 5. Residential inverter market in California from the California NEM database (Go Solar CA
2017), 2010–2016 .................................................................................................................... 7
Figure 6. Commercial inverter market in California from the California NEM database (Go Solar
CA 2017), 2010–2016 .............................................................................................................. 7
Figure 7. Non-MLPE inverter prices (USD/Wac) from PVinsights (2017), Q1 2017................................ 10
Figure 8. MLPE inverter shipments and prices (USD/Wac) from public corporate filings
(Enphase 2017, SolarEdge 2017), Q1 2014–Q1 2017 ........................................................... 11
Figure 9. Ex-factory gate price (spot prices) for U.S. crystalline-silicon modules from Bloomberg
(2017) data ............................................................................................................................. 12
Figure 10. Actual market module prices (2017 USD) ................................................................................ 13
Figure 11. Updated bottom-up manufacturing cost model results for the full crystalline-silicon module
supply chain from 2014/15 to Q1 2017.................................................................................. 15
Figure 12. Residential PV market share: integrator vs. installer, Q1 2014–Q1 2016 (GTM Research
and SEIA 2017; Sunrun 2017; Vivint Solar 2017) ................................................................ 17
Figure 13. Residential PV: model structure ................................................................................................ 18
Figure 14. Q1 2017 U.S. benchmark: 5.7-kW residential system cost (2017 USD/Wdc) .......................... 21
Figure 15. Q1 2017 benchmark by location: 5.7-kW residential system cost (2017 USD/Wdc) ............... 21
Figure 16. Q1 2017 NREL modeled cost benchmark (2017 USD/Wdc) vs. Q4 2016 company-
reported costs ......................................................................................................................... 22
Figure 17. NREL residential PV system cost benchmark summary (inflation adjusted), Q4 2009–Q1
2017........................................................................................................................................ 23
Figure 18. Levelized cost of energy for residential PV systems, by region and with and without ITC,
2010 –2017............................................................................................................................. 26
Figure 19. Commercial PV: model structure .............................................................................................. 27
Figure 20. Q1 2017 U.S. benchmark: commercial system cost (2017 USD/Wdc) ..................................... 29
Figure 21. Q1 2017 benchmark by location: 200-kW commercial system cost (2017 USD/Wdc) ............ 30
Figure 22. NREL commercial PV system cost benchmark summary (inflation adjusted), Q4 2009–
Q1 2017 .................................................................................................................................. 31
Figure 23. Levelized cost of energy for commercial PV systems, by region and with and without ITC,
2010 –2017............................................................................................................................. 33
Figure 24. Utility-scale PV: model structure .............................................................................................. 34
Figure 25. Percentage of U.S. utility-scale PV systems using tracking systems, 2007–2016 (Bolinger
and Seel 2017)........................................................................................................................ 36
Figure 26. Utility-scale PV: 2016 capacity installed and percentage of unionized labor by state
(BLS 2017; GTM Research and SEIA 2017) ........................................................................ 37
Figure 27. Q1 2017 benchmark by location: 100-MW utility-scale PV systems, EPC only (2017
USD/Wdc).............................................................................................................................. 38
Figure 28. Q1 2017 U.S. benchmark: utility-scale PV total cost (EPC + developer), 2017 USD/Wdc ..... 39

xi
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 29. NREL utility-scale PV system cost benchmark summary (inflation adjusted), 2010–2017 ..... 40
Figure 30. Levelized cost of energy for utility-scale PV systems, by region and with and without ITC,
2010–2017.............................................................................................................................. 43
Figure 31. Model application: U.S. utility-scale one-axis tracking PV system cost reduction from
economies of scale (2017 USD/Wdc) .................................................................................... 44
Figure 32. Modeled impacts of module efficiency on total system costs, 2017 ......................................... 45
Figure 33. Modeled real LCOE (¢/kWh), ITC = 0%, for a 100-MWdc utility-scale PV system with
fixed tilt and one-axis tracking in 2017.................................................................................. 47
Figure 34. NREL PV system cost benchmark summary (inflation adjusted), 2010–2017 ......................... 49
Figure 35. Modeled trend of soft cost as a proportion of total cost by sector, 2010–2017 ......................... 51
Figure 36. NREL PV LCOE benchmark summary (inflation adjusted), 2010–2017 ................................. 51

xii
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 List of Tables
Table ES-1. Benchmark Assumptions ......................................................................................................... iv
Table ES-2. Comparison of Q1 2016 and Q1 2017 PV System Cost Benchmarks .................................... vii
Table 1. Comparison of Inverter Solutions: String Inverter, DC Power Optimizer, and Microinverter ....... 6
Table 2. Rapid-Shutdown Codes—Progress by State ................................................................................... 8
Table 3. Rapid Shutdown—Different Inverter Solutions ............................................................................. 9
Table 4. Inverter Price Conversion (2017 USD)......................................................................................... 11
Table 5. Installer and Integrator Cost Changes, Q1 2016–Q1 2017 ........................................................... 17
Table 6. Residential PV: Modeling Inputs and Assumptions ..................................................................... 19
Table 7. Residential PV LCOE Assumptions, 2010–2017 ......................................................................... 24
Table 8. Commercial PV: Modeling Inputs and Assumptions ................................................................... 28
Table 9. Commercial PV LCOE Assumptions, 2010–2017 ....................................................................... 32
Table 10. Utility-Scale PV: Modeling Inputs and Assumptions ................................................................. 35
Table 11. One-Axis Tracker and Fixed-Tilt Utility-Scale PV LCOE Assumptions, 2010–2017 ............... 42
Table 12. Comparison of Q1 2016 and Q1 2017 PV System Cost Benchmarks ........................................ 50
Table 13. NREL Residential PV Benchmark Summary (Inflation Adjusted), 2010–2017 ........................ 54
Table 14. NREL Commercial PV Benchmark Summary (Inflation Adjusted), 2010–2017....................... 55
Table 15. NREL Utility-Scale PV Benchmark Summary (Inflation Adjusted), 2010–2017 ...................... 56
Table 16. NREL LCOE Summary (2017 cents/kWh) ................................................................................ 57

xiii
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 1 Introduction
Solar photovoltaic (PV) deployment has grown rapidly in the United States over the past several
years. As Figure 1 shows, in 2016 new U.S. PV installations included 2.3 gigawatts (GW) in the
residential sector, 1.1 GW in the commercial sector, and 10.2 GW in the utility-scale sector—
totaling 13.7 GW across all sectors (Bloomberg 2017). At the same time, PV system costs have
continued to decline. Previous modeling (Fu et al. 2016) by the National Renewable Energy
Laboratory (NREL) shows system cost reductions of about 60%–80% across sectors between
the fourth quarter of 2009 (Q4 2009) and Q1 2016.

Figure 1. U.S. PV market growth, 2004–2016, in gigawatts of direct-current (DC) capacity

(Bloomberg 2017)
This report continues tracking cost reductions by benchmarking costs of U.S. PV for residential,
commercial, and utility-scale systems built in Q1 2017. It was produced in conjunction with
several related research activities at NREL and Lawrence Berkeley National Laboratory, which
are documented in Barbose and Darghouth (2016), Bolinger and Seel (2016), Chung et al.
(2015), Feldman et al. (2015), and Fu et al. (2016).

Our methodology includes bottom-up accounting for all system and project-development costs
incurred when installing residential, commercial, and utility-scale systems, and it models the Q1
2017 costs for such systems excluding any previous supply agreements or contracts. In general,
we attempt to model the typical installation techniques and business operations from an installed-
cost perspective, and our benchmarks are national averages of installed capacities, weighted by
state. The residential benchmark is further averaged across installer and integrator business
models, weighted by market share. All benchmarks assume non-union construction labor,
although union labor cases are estimated for utility-scale systems.

1
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Our modeled costs can be interpreted as the sales price an engineering, procurement, and
construction (EPC) contractor/developer might charge for a system before any developer fee or
price gross-up. We use this approach owing to the wide variation in developer profits in all three
sectors, where project pricing is highly dependent on region and project specifics such as local
retail electricity rate structures, local rebate and incentive structures, competitive environment,
and overall project or deal structures.

The remainder of this report is organized as follows. Section 2 describes our model inputs
and sources. Sections 3, 4, and 5 show specific model inputs and outputs for the residential,
commercial, and utility-scale PV sectors, including historical trends in system costs and the
levelized costs of energy (LCOE). Section 6 includes three additional applications of our cost
modeling: system cost reduction from economies of scale, module efficiency impacts, and
regional LCOEs. Finally, Section 7 puts the results in context with each other and offers
conclusions.

2
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 2 Model Inputs and Sources
This section describes our model inputs and sources. Section 2.1 describes our main data source,
California’s Net Energy Metering (NEM) Interconnection Applications Data Set. Sections 2.2
through 2.6 detail the inputs for the various components affecting PV system cost, and Section
2.7 describes how we allocated installations to installers versus integrators in the residential PV
model.

2.1 California’s NEM Interconnection Applications Data Set

Previous NREL analyses used the California Solar Initiative Data Set (CSI 2017), but, as that
program has wound down, the number of new PV incentive applications—and consequently the
data collection—has decreased substantially. As a result, in last year’s report, we began using the
more robust California NEM Interconnection Applications Data Set instead (Go Solar CA 2017).
This database is updated monthly and contains all interconnection applications in the service
territories of the state’s three investor-owned utilities (Pacific Gas & Electric, Southern
California Edison, and San Diego Gas & Electric). We use the database to benchmark generic
system characteristics, such as system size, module power and efficiency, and choice of power
electronics. Although there are other databases for other markets, such as Massachusetts and
New York, we use only the California NEM database to inform these general benchmark
characteristics because of its higher granularity and greater consistency. Notably, we do not use
the California NEM database for regional cost analyses. Inputs and sources for regional analyses
are described in subsequent sections of this report.

As shown in Figure 2, the California NEM database captures most residential capacity in
California (79% of installed capacity in 2015 and 80% in 2016) and a sizable portion of
commercial capacity (91% of installed capacity in 2015 and 35% in 2016). Note that:

• We analyze only rooftop systems in the database for the residential and commercial
sectors. We exclude ground-mounted systems.
• We exclude systems with only alternating-current (AC) power records.
• We exclude systems that were still in the validation phase.
• We use GTM (2017) data to represent total installed capacities.

3
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 2. Installed capacities of residential and commercial PV systems covered by the California
NEM database (Go Solar CA 2017) compared with GTM data (GTM Research 2017), 2010–2016

2.2 Module Power and Efficiency

Figure 3 displays module power and efficiency data from the California NEM database. Since
2010, module power and efficiency in both sectors have been steadily improving. We use the
values of 16.2% (residential) and 17.5% (commercial and utility-scale) module efficiency in our
models. Also note that since module selection may vary in different regions, the actual module
efficiencies in other regions than CA may be different.

Figure 3. Module power and efficiency trends from the California NEM database
(Go Solar CA 2017), 2010–2016

4
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 2.3 PV System Size
Figure 4 displays average system sizes from the California NEM database. Average residential
system sizes have not changed significantly over the past 6 years. We use the 2016 value of
5.7 kW as the baseline case in our residential cost model. Conversely, commercial system sizes
have changed more frequently, likely reflecting the wide scope for “commercial customers,”
which include schools, office buildings, malls, retail stores, and government projects. We use
200 kW as the baseline case in our commercial model.

Figure 4. PV system size trends from the California NEM database (Go Solar CA 2017), 2010–2016

2.4 Module-Level Power Electronics

Microinverters and DC power optimizers are collectively referred to as module-level power
electronics (MLPE). By allowing designs with different roof configurations (orientations and
tilts) and constantly tracking the maximum power point for each module, MLPE provide an
optimized design solution at the module level. Table 1 provides a brief comparison of traditional
string inverters and MLPE.

5
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 1. Comparison of Inverter Solutions: String Inverter, DC Power Optimizer, and Microinverter

String Inverter DC Power Optimizer Microinverter

Each PV module has one
PV modules are
power optimizer for DC-to-
connected in parallel by
DC conversion, so the Each PV module has one
one or multiple strings
traditional junction box is microinverter for DC-to-
and then directly
replaced, and all modules AC conversion, and thus
Function connected to the string
are connected by string no string inverter is used.
inverter for DC-to-AC
inverter for DC-to-AC Shading only impacts
conversion. If one
conversion. Shading only individual modules.
module is shaded, the
impacts individual
whole string is impacted.
modules.
Low (without rapid
Relative shutdown)
Medium High
product price Medium (with rapid
shutdown)
Performance in
Poor More efficient More efficient
shading
Performance in
various
Low Medium High
directions or on
irregular roofs
Module-level
Yes (e.g., SolarEdge Yes (e.g., Enphase
monitoring and No
Cellular Kit) “Envoy + Enlighten”)
troubleshooting
Improved
energy yield
from module No Yes Yes
mismatch
reduction
Number of
Greater (thus may have Greater (thus may have
electronic Normal
some component risks) some component risks)
components
Safest; use only AC cable
Safety for
Normal Safer; easier wiring work with no high-voltage DC
installation
power

According to the California NEM database, market uptake of MLPE has been growing rapidly
since 2010 in California’s residential sector (Figure 5). This increasing market growth may be
driven by decreasing MLPE costs and by the “rapid shutdown” of PV output from buildings
required by Article 690.12 of the National Electric Code (NEC) since 2014—MLPE inherently
meet rapid-shutdown requirements without the need to install additional electrical equipment.

In 2016, MLPE—represented by the combined share of Enphase and SolarEdge inverter

solutions—reached 53% of the total California residential market share (Figure 5). Therefore, in
our residential system cost model, string inverter, power optimizer, and microinverter options are
modeled separately and their market shares (47%, 26%, and 27%) are used for the weighted
average case. Conversely, MLPE growth (represented by Enphase and SolarEdge) has been slow

6
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 in California’s commercial sector, reaching a share of only 12% in 2016 (Figure 6). Thus, we do
not include MLPE inverter solutions into our commercial model.

Figure 5. Residential inverter market in California from the California NEM database
5
(Go Solar CA 2017), 2010–2016

Figure 6. Commercial inverter market in California from the California NEM database
(Go Solar CA 2017), 2010–2016

5
“Others” represents other companies with small market shares. Although some companies may also have MLPE-
based inverter products, we assume that SolarEdge and Enphase represent MLPE inverter manufacturers.

7
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 For safety reasons, rapid-shutdown codes 6 are prevalent in most of the top residential PV
markets, and they typically include language from NEC 2014 (Article 690.12). 7 As of January 1,
2017, the 2017 NEC rapid-shutdown code was in effect in one state, the 2014 NEC was in effect
in 35 states, the 2011 NEC was in effect in five states, and the 2008 NEC was in effect in six
states (Table 2). Our cost model uses the 2014 NEC, which is the most widely adopted version
and includes the rapid-shutdown requirement. Table 3 presents the rapid-shutdown technical
solutions and cost impacts for various inverter options. Because of the increase in rapid
shutdown requirements, the cost difference between string inverter and power optimizer
configurations became smaller this year. 8 The model for our Q1 2016 benchmark did not include
rapid shutdown.

Table 2. Rapid-Shutdown Codes—Progress by State

Code Rapid-Shutdown State

Requirement

2017 NEC Yes Massachusetts

Alabama, Alaska, Arkansas, California, Colorado, Connecticut,

Delaware, Georgia, Idaho, Iowa, Kentucky, Maine, Maryland,
Michigan, Minnesota, Montana, Nebraska, New Hampshire, New
2014 NEC Yes Jersey, New Mexico, New York, North Carolina, North Dakota,
Ohio, Oklahoma, Oregon, Rhode Island, South Carolina, South
Dakota, Texas, Utah, Vermont, Washington, West Virginia,
Wyoming
2011 NEC No Florida, Louisiana, Virginia, Wisconsin, Nevada
2008 NEC No Hawaii, Illinois, Indiana, Kansas, Pennsylvania, Tennessee
No
statewide
No Arizona, Mississippi, Missouri
NEC
adoption

6
During a power shutdown (e.g., during a building fire or utility power loss), DC conductors in each PV array string
are most dangerous to first responders such as fire fighters because the DC side can still be energized even if the
inverter is shut down. Rapid-shutdown codes require a set distance between PV system conductors and PV arrays, so
the conductors are de-energized to a safe level and risks to first responders are reduced.
7
For example, a segment of the NEC language that is used says, “Conductors more than 5 feet inside a building or
more than 10 feet from an array will be limited to a maximum of 30 V and 240 VA within 10 seconds of shutdown.”
This only applies to PV system circuits “on or in buildings,” thus ground-mounted systems are not required to have
rapid-shutdown capability.
8
The costs were $2.78/W (string inverter) vs. $2.94/W (power optimizer) in Q1 2016 when rapid shutdown was not
included in our cost models, compared with $2.90/W (string inverter) vs. $2.95/W (power optimizer) if rapid
shutdown is included in Q1 2016 benchmark.

8
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 3. Rapid Shutdown—Different Inverter Solutions

String Inverter DC Power Optimizer Microinverter

A rapid-shutdown box
must be mounted Microinverters inherently meet
directly to the PV rapid-shutdown requirements
mounting rail and fit A rapid-shutdown cable
Solution for without any additional electrical
under the PV modules. must be installed in the
rapid- equipment, because the DC side
inverter box. No additional
shutdown A rapid-shutdown (which has low voltage) is de-
roof-mounted devices are
requirement controller must be energized as soon as the grid or
required.
mounted so it is visible power from the grid is
and freely accessible to interrupted.
first responders.
Rapid shutdown box
Additional Rapid shutdown
One rapid shutdown cable
balance-of- controller
in each inverter
system (BOS) Cable between box and None
costs Total BOS increase =
controller
$0.01/W
Total BOS increase =
$0.08/W
Electrician for cabling
between box and
controller Electrician for setting up
Additional
internal cable in each
direct labor Common labor for inverter None
costs racking box and
controller Total labor increase =
$0.01/W
Total labor increase =
$0.01/W
Q1 2016 –
Benchmark
(no rapid $2.78/W $2.94/W $3.28/W
shutdown
consideration)
Q1 2016 –
Benchmark (if
rapid $2.90/W $2.95/W $3.28/W
shutdown is
considered)
Cost change
0.12/W = 0.08/W
in 2016
(electrical BOS) +
models due to 0.01/W = 0.01/W (electrical
0.01/W (direct labor) + No change
rapid BOS and direct labor)
0.03/W (other related
shutdown
costs)
only

9
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 2.5 Inverter Price and DC-to-AC Ratios
As shown in Figure 7, we source non-MLPE inverter prices from the PVinsights (2017)
database, which contains typical prices between Tier 1 suppliers and developers in the market.
For MLPE inverter prices, we use data from public corporate filings, shown in Figure 8 (Enphase
2017; SolarEdge 2017). 9 Enphase’s Q1 2017 revenue was $0.40/Wac, which represents the
typical microinverter price. SolarEdge’s Q1 2017 revenue was $0.25/Wac, including sales from
DC power optimizers, string inverters, and monitoring equipment, which are typically included
in one product offering. GTM Research estimates a DC power optimizer cost of $0.08/Wac
(GTM Research 2017), implying a string inverter and monitoring equipment price of $0.17/Wac.
This is close to the Q1 2017 non-MLPE string inverter costs of $0.15/Wac shown in Figure 7
(assuming a $0.02–$0.03/Wac cost for monitoring equipment) (GTM Research and SEIA 2017).

We convert the USD/Wac inverter prices from Figure 7 and Figure 8 to USD per watt DC (Wdc)
using the DC-to-AC ratios shown in Table 4. In our benchmark, we use USD/Wdc for all costs,
including inverter prices.

Figure 7. Non-MLPE inverter prices (USD/Wac) from PVinsights (2017), Q1 2017

9
All sourced inverter prices are quoted in U.S. dollars (USD) per watt AC (Wac).

10
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 8. MLPE inverter shipments and prices (USD/Wac) from public corporate filings
(Enphase 2017, SolarEdge 2017), Q1 2014–Q1 2017

Table 4. Inverter Price Conversion (2017 USD)

a
Inverter Type Sector USD/Wac DC-to-AC Ratio USD/Wdc

Single-Phase String Residential PV (non-

0.15 1.15 0.13
Inverter MLPE)

Residential PV
Microinverter 0.40 1.15 0.34
(MLPE)

DC Power Optimizer Residential PV

0.17 1.15 0.15
String Inverter (MLPE)

Three-Phase String Commercial PV (non-

0.12 1.15 0.10
Inverter MLPE)

Utility-scale PV (fixed- b
Central Inverter 0.08 1.3 (oversized) 0.06
tilt)

Utility-scale PV (1-
Central Inverter 0.08 1.3 (oversized) 0.06
axis tracker)
a
We updated the central inverter DC-to-AC ratios using Lawrence Berkeley National Laboratory data
(Bolinger and Seel 2017); for the other ratios, we use the estimates from our 2016 report (Fu et al. 2016)
based on interview feedback (NREL 2017).
b
A DC-to-AC ratio larger than one means that the PV array’s DC rating is higher than the inverter’s AC
rating. This increases inverter utilization, although it also results in some PV energy curtailment, or
“clipping,” during the sunniest periods when PV output exceeds the inverter’s capacity. PV module prices

11
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 have dropped more rapidly than inverter prices have, and many utility-scale PV developers have found it
economical to oversize their PV arrays. The resulting AC-generation gains during periods of less-than-
peak PV production more than offset the losses from occasional peak-period clipping (Bolinger and Seel
2016).

2.6 Module Prices

We use $0.35/W—the spot price of U.S. crystalline-silicon modules in March 2017—to
represent the ex-factory gate price between Tier 1 module suppliers and first buyers 10 in all
sectors, based on Bloomberg (2017) data (Figure 9). Because we model ex-factory gate price in
Q1 2017, actual market pricing may vary owing to previously signed supply agreements or
installer/distributor inventory lags. 11 In addition, the actual market price may vary by market
segment because of increased supply-chain costs as well as the price premium for small-scale
procurement. Compared with module spot prices in 2016, module spot prices in 2017 have also
been influenced by changes in currency exchange rates. The USD appreciated against the
Chinese Yuan by approximately 6% between Q1 2016 and Q1 2017 (XE Currency Charts 2017).

Figure 9. Ex-factory gate price (spot prices) for U.S. crystalline-silicon modules from
Bloomberg (2017) data
Despite a $0.35/W factory gate module price, additional module costs increase national
integrators’ total module costs to $0.65/W (86% price premium) and small installers’ total
module costs to $0.73/W (109% price premium). These additional costs in Figure 10 consist of
shipping and handling (a 15% price premium above factory gate pricing for national integrators
and small installers, respectively [NREL 2017]), historical inventory (a 60% price premium

10
The first buyers of modules ex-factory gate can be developers, EPC contractors, installers, distributors, retailers,
or other end users. In our cost model, first buyer price—that is, ex-factory gate price—is used as the “module price”
component of the total system cost in the residential, commercial, and utility-scale sectors.
11
The effect of inventory lags and previous supply agreements on system pricing in the latter half of 2016 and the
first quarter of 2017 may be particularly high, because the actual market module price had not dropped so
precipitously since 2011 and 2012.

12
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 above factory gate pricing [NREL 2017]), a sales-tax of 6.7%, and, for small installers, a 20%
price premium above factory gate pricing due to small-scale procurement (Bloomberg 2017).

In Q1 2017 historical inventory represented the largest supply-chain cost for residential
installers. While we do not include pre-existing supply agreements or other contracts into our
benchmark, historical inventory is a necessary cost for residential installers. Because
homeowners of residential rooftop PV systems have different preferences for module brand, both
small installers and national integrators tend to diversify their module procurement. Furthermore,
since rooftop PV system sizes are relatively small (5.7 kW in our benchmark), the various
module brands procured may not be fully consumed and installed instantly. Thus, the historical
inventory price creates a price lag (approximately six months) for the market module price in the
residential sector when the modules from previous procurement are installed in today’s systems.

From 2012 to mid-2016 this price lag did not create a large price premium because the average
spot price of modules did not change dramatically. However, from mid-2016 to early-2017
module spot price dropped by approximately $0.25/W, or 41%, as shown in Figure 9. Thus, in
the first quarter of 2017 residential installers must bear the costs of this $0.21/W historical
inventory. It is likely that this price premium will be much smaller next year as analysts expect
the spot price curve to become flatter. However, many things may change within the market
(e.g., tariffs) and make it challenging for residential players to forecast module price. Without
historical inventory, total module costs would be $0.43/W for national integrators and $0.52/W
for small installers (potentially reducing total residential PV costs to $2.59/Wdc).

Figure 10. Actual market module prices (2017 USD)

Besides module spot price, actual module manufacturing cost is introduced here in order to
demonstrate the technology improvement. We work across the spectrum of academic and
national laboratory researchers, startup companies, and multinational corporations to understand
the cost drivers and technology landscape of PV module production. Our bottom-up method
entails an examination of each stage in the supply chain, including polysilicon, ingot, and wafer
production, cell conversion, and module assembly. For each stage, we begin with the derivation

13
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 of detailed technology-manufacturing process flows. Then we work with equipment and
materials suppliers, as well as integrated manufacturers already engaged in production, to collect
and verify the costs for each step of the process. Finally, we sum the individual process steps to
generate total costs for the intermediate materials (polysilicon, ingots, wafers, and cells) and
finished PV modules.

Figure 11 shows our most recent module manufacturing cost analysis, for passivated emitter and
rear cells (PERC) and modules manufactured in Southeast Asia. The dark blue bars show the Q1
2017 cost contributions for each step: about $0.05/W for polysilicon, $0.05/W for ingot and
wafer production, $0.08/W for cell conversion, $0.13/W for module assembly, and $0.03/W for
an industry-average budget for research and development (R&D) plus sales, general, and
administrative (SG&A). The all-in module manufacturing cost is about $0.35/W.

Figure 11 also illustrates the magnitude of cost reductions since our last detailed module
manufacturing analysis in 2014 and the first half of 2015, when we calculated an all-in module
manufacturing cost of about $0.63/W. This 45% reduction in costs over 2–3 years was enabled
by improving silicon utilization (principally reducing kerf loss), converting from slurry-based
wafer slicing to diamond-wire-based wafer slicing, and reducing costs for cell conversion and
module assembly principally via improved efficiency and capital investment requirements (the
depreciation expenses shown in the figure). In a forthcoming paper, we will detail additional
technology-improvement opportunities that could lead to even lower costs in the future.

14
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 12
Figure 11. Updated bottom-up manufacturing cost model results for the full crystalline-silicon module supply chain from 2014/15 to Q1 2017

12
The results shown are for manufacturing PERC and modules in Southeast Asia.
15
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 2.7 Small Installers vs. National Integrators in the Residential
PV Model
Our residential PV benchmark is based on two different business structures: “small installer” and
“national integrator.” We define small installers as businesses that engage in lead generation,
sales, and installation, but do not provide financing solutions. The national integrator performs
all of the small installer’s functions, and provides financing and system monitoring for third-
party-owned systems. In our models, the difference between small installers and national
integrators is manifested in the overhead and sales and marketing cost categories, where the
national integrator is modeled with higher expenses for customer acquisition, financial
structuring, and asset management.

To estimate the split in market share between small installers and national integrators, we use
data compiled from corporate filings (Sunrun 2017; Vivint Solar 2017) and GTM Research and
SEIA (2017). As shown in Figure 12, small installers gained more market share than national
integrators did during 2016, in part because the direct ownership business model, led by
installers, remained more popular than third-party ownership. We use the 41% integrator and
59% installer market shares in our Q1 2017 model to compute the national weighted-average
case in our residential PV model.

Table 5 summarizes overhead and sales and marketing costs for small installers and national
integrators from our Q1 2016 and Q1 2017 reports. National integrators achieved lower per-watt
sales and marketing and overhead costs in Q1 2017 compared with Q1 2016 because of lower
reported total expenditures on those two categories. Small installers had higher total expenditures
on sales and marketing and overhead as they prepared to grow their businesses in 2017, but they
still achieved lower per-watt costs for sales and marketing in Q1 2017 compared with Q1 2016
because they installed more PV capacity in the later period.

16
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 12. Residential PV market share: integrator vs. installer, Q1 2014–Q1 2016 (GTM Research
and SEIA 2017; Sunrun 2017; Vivint Solar 2017)

Table 5. Installer and Integrator Cost Changes, Q1 2016–Q1 2017

Q1 2016 Report Q1 2017 Report

Sales & marketing $0.31/Wdc (small installer) $0.29/Wdc (small installer)

(customer
acquisition) $0.43/Wdc (national integrator) $0.42/Wdc (national integrator)

Overhead (general $0.28/Wdc (small installer) $0.28/Wdc (small installer)

& administrative) $0.38/Wdc (national integrator) $0.35/Wdc (national integrator)

17
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 3 Residential PV Model
This section describes our residential model’s structure, inputs, and assumptions (Section 3.1),
output (Section 3.2), and differences between modeled output and reported costs (Section 3.3).

3.1 Residential Model Structure, Inputs, and Assumptions

We model a 5.7-kW residential rooftop system using 60-cell, multicrystalline, 16.2%-efficient
modules from a Tier 1 supplier and a standard flush mount, pitched-roof racking system. Figure
13 presents the cost drivers and assumptions, cost categories, inputs, and outputs of the model.
Table 6 presents modeling inputs and assumptions in detail.

Figure 13. Residential PV: model structure

18
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 6. Residential PV: Modeling Inputs and Assumptions

Category Modeled Value Description Sources

System size 5.7 kW Average installed size per system Go Solar CA (2017)
Module
16.2% Average module efficiency Go Solar CA (2017)
efficiency
Ex-factory gate (first buyer) price, Tier 1 Bloomberg (2017), NREL
Module price $0.35/Wdc
modules (2017)
Single-phase
string inverter:
$0.13/Wdc
Go Solar CA (2017), NREL
DC power Ex-factory gate (first buyer) prices, Tier 1 (2017), PVinsights (2017),
Inverter price optimizer string inverters corporate filings (Enphase
inverter: 2017; SolarEdge 2017)
$0.15/Wdc
Microinverter:
$0.34/Wdc
Structural BOS Model assumptions, NREL
$0.11/Wdc Includes flashing for roof penetrations
(racking) (2017)
$0.20– Conductors, switches, combiners and
$0.33/Wdc transition boxes, as well as conduit, Model assumptions, NREL
Electrical BOS
Varies by grounding equipment, monitoring system or (2017), RSMeans (2016)
inverter option production meters, fuses, and breakers
15% costs and fees associated with
shipping and handling of equipment
multiplied by the cost of doing business
index (101%)
Supply chain Additional 80% (60% historical inventory +
costs (% of Varies by 20% small-scale procurement) for module- NREL (2017), model
equipment installer type related supply chain costs for small assumptions (2017)
costs) installers and 60% (historical inventory) for
national integrators
Additional 20% for inverter-related supply
chain costs for small installers and 10% for
national integrators
Sales tax on the equipment; national
Varies by DSIRE (2017), RSMeans
Sales tax benchmark applies an average (by state)
location (2016)
weighted by 2016 installed capacities
Electrician:
$19.37–$38.22
per hour;
Direct Laborer: Modeled labor rate depends on state;
installation $12.64–$25.09 national benchmark uses weighted BLS (2017), NREL (2017)
labor per hour; average of state rates
Varies by
location and
inverter option

19
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Category Modeled Value Description Sources
Workers compensation (state-weighted
Burden rates Total
average), federal and state unemployment
(% of direct nationwide RSMeans (2016)
insurance, Federal Insurance Contributions
labor) average: 31.8%
Act (FICA), builders risk, public liability
Includes assumed building permitting fee of
Permitting,
$400 and six office staff hours for building NREL (2017), Vote Solar
inspection, and
$0.10/Wdc permit preparation and submission, and (2015), Vote Solar and
interconnection
interconnection application preparation and IREC (2013)
(PII)
submission
Total cost of sales and marketing activities
Sales & $0.29/Wdc over the last year—including marketing and
(installer) NREL (2017), Sunrun
marketing advertising, sales calls, site visits, bid
(2017), Vivint Solar (2017),
(customer $0.42/Wdc preparation, and contract negotiation;
Feldman et al. (2013)
acquisition) (integrator) adjusted based on state “cost of doing
business” index
General and administrative expenses—
including fixed overhead expenses
$0.28/Wdc covering payroll (excluding permitting
Overhead (installer) NREL (2017), Sunrun
payroll), facilities, administrative, finance,
(general & (2017), Vivint Solar (2017),
$0.35/Wdc legal, information technology, and other
administrative) Feldman et al. (2013)
(integrator) corporate functions as well as office
expenses; adjusted based on state “cost of
doing business” index
Applies a fixed percentage margin to all
direct costs including hardware, installation
Profit (%) 17% Fu et al. (2016)
labor, direct sales and marketing, design,
installation, and permitting fees

3.2 Residential Model Output

Figure 14 presents the U.S. national benchmark from our residential model. The national
benchmark represents an average weighted by 2016 state installed capacities. Market shares of
59% for installers and 41% for integrators are used to compute the national weighted average.
String inverter, power optimizer, and microinverter options are each modeled individually, and
the “mixed” case applies their market shares (47%, 26%, and 27%) 13 as weightings.

Small installers have lower total costs than do large integrators; although small installers pay
more for hardware, they have much lower overhead and sales and marketing costs. Notably, the
cost difference between installer and integrator became smaller in Q1 2017 than in Q1 2016 (see
Table 5). Because of rapid-shutdown requirements, the cost difference between string inverters
and power optimizers also became smaller in Q1 2017 than in Q1 2016 (see Table 3).

13
This market share combination only reflects the California residential sector and may not reflect the actual
national market shares.

20
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 14. Q1 2017 U.S. benchmark: 5.7-kW residential system cost (2017 USD/Wdc)
Figure 15 presents the benchmark in the top U.S. solar markets (by 2016 installations), reflecting
differences in supply chain and labor costs, sales tax, and SG&A expenses—that is, the cost of
doing business (Case 2012).

Figure 15. Q1 2017 benchmark by location: 5.7-kW residential system cost (2017 USD/Wdc)

21
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 3.3 Residential Model Output vs. Reported Costs
As shown in Figure 16, our bottom-up modeling approach yields a different cost structure than
those reported by public solar integrators in their corporate filings 14 (Sunrun 2017; Vivint Solar
2017). Because integrators sell and lease PV systems, they practice a different method of
reporting costs than do businesses that only sell goods. Many of the costs for leased systems are
reported over the life of the lease rather than the period in which the system is sold; therefore, it
is difficult to determine the actual costs at the time of the sale. Although there are the corporate
filings from Sunrun and Vivint Solar report system costs on a quarterly basis, the limited
transparency in the public filings makes it difficult to determine the underlying costs as well as
the timing of those costs. As indicated in Figure 16, our total modeled costs for national
integrators are $0.40–$0.46/W below company-reported values. Because of the lack of
transparency in the reported company costs, it is difficult to explain these differences entirely.
Part of the difference in installation costs could come from integrators having preexisting
contracts or older inventory that they used in systems installed in Q1 2017; this is particularly
relevant owing to the rapid decline in module price in the second half of 2016. In addition, our
sales and marketing costs are $0.08–$0.23/W below company-reported values, indicating either a
difference in how costs are classified or additional costs not included in our model—a deeper
exploration of this topic may prove valuable.

Figure 16. Q1 2017 NREL modeled cost benchmark (2017 USD/Wdc) vs. Q4 2016 company-
reported costs

14
Because of the acquisition of SolarCity by Tesla, the quarterly corporate filings from SolarCity are not available
this year.

22
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 3.4 Residential PV Price Benchmark Historical Trends
NREL began benchmarking PV system costs in 2010 in order to track PV system energy costs
against the U.S. Department of Energy’s (DOE) SunShot Initiative targets, as well as examine
cost reduction opportunities for achieving these goals. 15 Since that time NREL has produced
seven additional benchmarks, including a historical Q4 2009 benchmark. Figure 17 summarizes
the reduction in residential PV system cost benchmarks between 2010 and 2017. 16

Figure 17. NREL residential PV system cost benchmark summary (inflation adjusted),
Q4 2009–Q1 2017

As demonstrated in Figure 17, from 2010 to 2017 there was a 61% reduction in the residential
PV system cost benchmark. Approximately 61% of that reduction can be attributed to total
hardware costs (module, inverter, and hardware BOS), as module prices dropped 86% over that
time period. An additional 18% can be attributed to labor, which dropped 73% over that time
period, with the final 21% attributable to other soft costs, including PII, sales tax, overhead, and
net profit.

Looking at this past year, from 2016 to 2017 there was a 6% reduction in the residential PV
system cost benchmark. The majority of that reduction can be attributed to the 46% reduction in
module factory gate price, moderated by the increase in module supply chain costs discussed
earlier (shown here in “soft costs – other”).

15
The original overarching 2020 goal of the SunShot Initiative was for solar to reach cost parity with baseload
energy rates, estimated to be 6 cents/kWh without subsidies, or a system installed cost of $1/W. Commercial PV and
residential PV were later separated to have their own goals of costs below retail rates, estimated to be 7 cents/kWh
and 9 cents/kWh respectively, or system installed costs of $1.25/W and $1.50/W respectively (note: all 2020 targets
are quoted in nominal USD). In recognition of the transformative solar progress to date and the potential for further
innovation, in 2016 the SunShot Initiative extended its goals to reduce the unsubsidized cost of energy by 2030 to
3¢/kWh, 4¢/kWh and 5¢/kWh for utility-scale PV, commercial PV, and residential PV (note: all 2030 targets are
quoted in nominal USD).
16
Each year’s PV system cost benchmark corresponds to the NREL benchmark calculted in Q4 of the previous year
or Q1 of the current year (e.g. 2010 = Q4 2009; 2017 = Q1 2017).

23
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 3.5 Residential PV Levelized Cost of Energy Historical Trends
While LCOE is not a perfect metric to measure the competiveness of PV within the energy
marketplace, it incorporates many other PV metrics important to the energy costs beyond upfront
installation costs. These benchmarks are summarized over time in Table 7, from Q4 2009 to Q1
2017 (aSunShot Vision Study 2010, bOn the Path to SunShot: The Role of Advancements in
Solar Photovoltaic Efficiency, Reliability, and Costs; cOn the Path to SunShot: Emerging
Opportunities and Challenges in Financing Solar (Feldman and Bolinger 2016); dTerms, Trends,
and Insights PV Project Finance in the United States (Feldman, Lowder and Schwabe 2016),
e
U.S. Solar Photovoltaic System Cost Benchmark: Q1 2016). 17

Table 7. Residential PV LCOE Assumptions, 2010–2017

2017 USD 2010 2011 2012 2013 2014 2015 2016 2017
per Watt DC
Installed cost $7.24 $6.34 $4.48 $3.92 $3.44 $3.18 $2.98 $2.80
a c
Annual 1.00% 0.95% 0.90% 0.85% 0.80% 0.75% 0.75% 0.75%
degradation
(%)
a c e
Inverter $0.41 $0.36 $0.31 $0.26 $0.21 $0.15 $0.14 $0.13
replacement
price ($/W)
a c
Inverter 10 11 12 13 14 15 15 15
lifetime
(years)
a c
O&M $37 $33 $30 $27 $24 $21 $21 $21
expenses
($/kw-yr)
a c
Pre-inverter 90.0% 90.10% 90.20% 90.30% 90.40% 90.5% 90.5% 90.5%
derate (%)
a c
Inverter 94.0% 94.80% 95.60% 96.40% 97.20% 98.0% 98.0% 98.0%
efficiency (%)
a c e
System size 5.0 5.0 5.1 5.1 5.2 5.2 5.6 5.7
(kw-DC)
a e
Inverter 1.1 1.11 1.12 1.13 1.13 1.14 1.15 1.15
loading ratio
c d
Equity 9.0% 8.6% 8.3% 7.9% 7.6% 7.3% 6.9% 6.9%
discount rate
e
(real)
a
Inflation rate 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5%
c d
Debt interest 5.5% 5.4% 5.3% 5.2% 5.0% 4.9% 4.8% 4.8%
f
rate
b d
Debt fraction 34.2% 35.2% 36.1% 37.1% 38.1% 39.0% 40.0% 40.0%

17
In instances in which LCOE assumptions were not found from the selected literature in a given year, straight-line
changes were assumed between any two values.

24
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 b
Other important assumptions: residential PV system LCOE assume a 1) system lifetime of 30 years , 2)
b b
federal tax rate of 35% , 3) state tax rate of 7% , 4) MACRS depreciation schedule, 5) no state or local
subsidies, 6) a working capital and debt service reserve account for six months of operating costs and
b
debt payments (earning an interest of 1.75%) , 7) a three month construction loan, with an interest rate of
b
4% and a fee of 1% of the cost of the system , 8) a module tilt angle of 25 degrees, and an azimuth of
b
180 degrees, 9) debt with a term of 18 years , and 10) $1.1MM of upfront financial transaction costs for a
d
$100MM TPO transaction of a pool of residential projects .
e
In instances in which LCOE assumptions were not found from the selected literature in a given year,
straight-line changes were assumed between any two values.
f
The historical financial structure for a residential TPO system assumed in 2010 from Feldman and
Bolinger 2016 does not assume a debt raise; however, the financial structure in 2016 from Feldman,
Lowder, and Schwabe does assume back-leveraged debt. To make these assumptions uniform, the “debt
interest rate” and “debt fraction” are taken from the utility-scale historical financial structure in Feldman
and Bolinger 2016 that uses back-leveraged debt.

As demonstrated in Table 7, in addition to a 61% reduction in installed cost from 2010 to 2017,
inverter replacement costs reduced 69%, O&M costs reduced 44%,annual degradation rates
reduced 25%, the equity discount rate reduced 23%, the debt interest rate reducd 13%, and the
debt fraction increased 17%.

Using these assumptions we calculated the LCOE, with and without the 30% federal investment
tax credit (ITC), in Phoenix, AZ, Kansas City, MO, and New York, NY, corresponding to
higher, medium, and lower resource areas in the United States and the locations used to calculate
LCOE in the SunShot Vision Study. The calculated values are summarized in Figure 18. 18

18
Because this analysis uses a more robust set of current and historical assumptions LCOE values may differ from
previously reported benchmarked values.

25
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 18. Levelized cost of energy for residential PV systems, by region and with and
without ITC, 2010 –2017
As demonstrated in Figure 18, from 2010 to 2017 there was a 70% reduction in the residential
PV system electricity cost benchmark (a 5% to 6% reduction was achieved from Q1 2016 to Q1
2017), bringing the unsubsidized LCOE between $0.13/kWh to $0.17/kWh ($0.08/kWh to
$0.11/kWh when including the federal ITC). This reduction is 86% toward achieving SunShot’s
2020 residential PV LCOE goal. 19

19
The SunShot 2020 target is adjusted from 2010 USD using the Consumer Price Index (CPI). A Summary of
these values can be found in Appendix A and B. For LCOE Kansas City, MO, without ITC cases are $0.52/kWh in
2010 and $0.16/kWh in 2017 in 2017 USD from Appendix A and B. Thus, calculation is: (0.52 – 0.16)/(0.52 – 0.10)
= 86%.

26
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 4 Commercial PV Model
This section describes our commercial model’s structure, inputs, and assumptions (Section 4.1)
and output (Section 4.2).

4.1 Commercial Model Structure, Inputs, and Assumptions

We model a 200-kW, 1,000 volts DC (Vdc), commercial-scale flat-roof system using
multicrystalline 17.5%-efficient modules from a Tier 1 supplier, three-phase string inverters, and
a ballasted racking solution on a membrane roof. A penetrating PV mounting system can have
higher energy yield (kWh per kW) owing to wider tilt-angle range allowance. However, we do
not model this system type, because its market share has declined owing to additional required
flashing and sealing work, roof warranty issues, and the relative difficulty of replacing such a
system in the future. Figure 19 presents a schematic of our commercial-scale system cost model.
Table 8 presents the detailed modeling inputs and assumptions. We separate our cost estimate
into EPC and project-development functions. Although some firms engage in both activities in
an integrated manner, and potentially achieve lower cost and pricing by reducing the total margin
across functions, we believe the distinction can help separate and highlight the specific cost
trends and drivers associated with each function.

Figure 19. Commercial PV: model structure

27
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 8. Commercial PV: Modeling Inputs and Assumptions

Category Modeled Value Description Sources

System size 10 kW – 2 MW Average installed size per system Go Solar CA (2017)
Module
17.5% Average module efficiency Go Solar CA (2017)
efficiency
Ex-factory gate (first buyer) price, Tier 1 Bloomberg (2017), NREL
Module price $0.35/Wdc
modules (2017)
Three-phase string
inverter: Ex-factory gate prices (first buyer) price, Bloomberg (2017), NREL
Inverter price
Tier 1 inverters (2017)
$0.10/Wdc
Structural $0.13–$0.28/Wdc;
ASCE (2006), model
components varies by location and Flat-roof ballasted racking system
assumptions, NREL (2017)
(racking) system size
Electrical Varies by location Conductors, conduit and fittings, transition Model assumptions, NREL
components and system size boxes, switchgear, panel boards, etc. (2017), RSMeans (2016)
EPC overhead
Costs and fees associated with EPC
(% of
13% overhead, inventory, shipping, and NREL (2017)
equipment
handling
costs)
Sales tax on equipment costs; national
DSIRE (2017), RSMeans
Sales tax Varies by location benchmark applies an average (by state)
(2016)
weighted by 2016 installed capacities
Electrician: $19.37–
$38.22 per hour Modeled labor rate assumes non-union
Direct
Laborer: $12.64– labor and depends on state; national
installation BLS (2017), NREL (2017)
$25.09 per hour benchmark uses weighted average of
labor
Varies by location state rates
and inverter option
Workers compensation (state-weighted
Burden rates
Total nationwide average), federal and state unemployment
(% of direct RSMeans (2016)
average: 31.8% insurance, FICA, builders risk, public
labor)
liability
For construction permits fee,
PII $0.11–$0.16/Wdc interconnection, testing, and NREL (2017)
commissioning
Assume 10-MW Includes fixed overhead expenses such
system development as payroll, facilities, travel, insurance,
Developer Model assumptions, NREL
and installation per administrative, business development,
overhead (2017)
year for a typical finance, and other corporate functions;
developer assumes 10 MW/year of system sales
Estimated as markup on EPC price; value
Contingency 4% represents actual cost overruns above NREL (2017)
estimated price
Applies a fixed percentage margin to all
costs including hardware, installation
Profit 7% NREL (2017)
labor, EPC overhead, developer
overhead, etc.

28
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 4.2 Commercial Model Output
Figure 20 presents the U.S. national benchmark from our commercial model. As in the
residential model, the national benchmark represents an average weighted by 2016 state-installed
capacities. We model different system sizes because of the wide scope of the “commercial”
sector, which comprises a diverse customer base occupying a variety of building sizes.
Economies of scale—driven by hardware, labor, and related markups—are evident here. As
system sizes increase, the per-watt cost to build them decreases. This holds even as we assume
that a typical developer has 10 MW of system development and installation per year, and
therefore has overhead on this 10 MW total capacity that does not vary for different system sizes.
When a developer installs more capacity annually, the developer’s overhead per watt in each
system declines (shown in Figure 18 in our Q1 2015 benchmark report, Chung et al. 2015).

Figure 20. Q1 2017 U.S. benchmark: commercial system cost (2017 USD/Wdc)
The PII cost was higher in Q1 2017 than in Q1 2016, because the low-hanging fruit—such as
ideal commercial building rooftops—have already been picked by Q1 2017. Thus, the associated
PII time and fees were higher in Q1 2017 for commercial projects with more PII obstacles. Also,
the higher net profit in Q1 2017—7%, compared with 2% in Q1 2016—indicates that the rapid
module price reduction in 2016 enabled EPC firms and developers to retain a higher profit and
still maintain a competitive project cost (NREL 2017).

29
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 21 presents the benchmark from our commercial model by location in the top U.S. solar
markets (by 2016 installations). The main cost drivers for different regions in the commercial PV
market are the same as in the residential model (labor rates, sales tax, and cost of doing business
index), but also include costs associated with wind or snow loading.

Figure 21. Q1 2017 benchmark by location: 200-kW commercial system cost (2017 USD/Wdc)

30
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 4.3 Commercial PV Price Benchmark Historical Trends
Figure 22 summarizes the reduction in commercial PV system cost benchmarks between 2010
and 2017. 20

Figure 22. NREL commercial PV system cost benchmark summary (inflation adjusted),
Q4 2009–Q1 2017
As demonstrated in Figure 22, from 2010 to 2017 there was a 65% reduction in the commercial
PV system cost benchmark. Approximately 82% of that reduction can be attributed to total
hardware costs (module, inverter, and hardware BOS), as module prices dropped 86% over that
time period. An additional 4% can be attributed to labor, which dropped 47% over that time
period, with the final 14% attibitubal to other soft costs, including PII, sales tax, overhead, and
net profit.

Looking at this past year, from 2016 to 2017 there was a 15% reduction in the commercial PV
system cost benchmark. The majority of that reduction can be attributed to the 46% reduction in
module factory gate price, moderated by an increase in PII and installer profit.

4.4 Commercial PV Levelized Cost of Energy Historical Trends

While LCOE is not a perfect metric to measure the competiveness of PV within the energy
marketplace, it incorporates many other PV metrics important to the energy costs beyond upfront
installation costs. These benchmarks are summarized over time in Table 9, from 2010 to 2017
(aSunShot Vision Study 2010, bOn the Path to SunShot: The Role of Advancements in Solar
Photovoltaic Efficiency, Reliability, and Costs; cOn the Path to SunShot: Emerging
Opportunities and Challenges in Financing Solar (Feldman and Bolinger 2016); dTerms, Trends,
and Insights PV Project Finance in the United States (Feldman, Lowder and Schwabe 2016),
e
U.S. Solar Photovoltaic System Cost Benchmark: Q1 2016). 21

20
Each year’s PV system cost benchmark corresponds to the NREL benchmark calculted in Q4 of the previous year
or Q1 of the current year (e.g. 2010 = Q4 2009; 2017 = Q1 2017).
21
In instances in which LCOE assumptions were not found from the selected literature in a given year, straight-line
changes were assumed between any two values.

31
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 9. Commercial PV LCOE Assumptions, 2010–2017

2017 USD 2010 2011 2012 2013 2014 2015 2016 2017
per Watt DC
Installed cost $5.36 $4.97 $3.42 $2.78 $2.76 $2.27 $2.17 $1.85
a b
Annual degradation 1.00% 0.95% 0.90% 0.85% 0.80% 0.75% 0.75% 0.75%
(%)
a b e
Inverter replacement $0.24 $0.22 $0.19 $0.17 $0.15 $0.12 $0.11 $0.10
price ($/W)
a b
O&M expenses $26 $24 $22 $20 $18 $15 $15 $15
($/kw-yr)
a b
Pre-inverter derate 90.5% 90.50% 90.50% 90.50% 90.50% 90.5% 90.5% 90.5%
(%)
a b
Inverter efficiency 95.0% 95.60% 96.20% 96.80% 97.40% 98.0% 98.0% 98.0%
(%)
a e
Inverter loading ratio 1.10 1.11 1.12 1.13 1.13 1.14 1.15 1.15
e c d
Equity discount rate 9.0% 8.6% 8.3% 7.9% 7.6% 7.3% 6.9% 6.9%
(real)
a
Inflation rate 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5%
f c d
Debt interest rate 5.5% 5.4% 5.3% 5.2% 5.0% 4.9% 4.8% 4.8%
c d
Debt fraction 34.2% 35.2% 36.1% 37.1% 38.1% 39.0% 40.0% 40.0%
b
Other important assumptions: commercial PV system LCOE assume a 1) system lifetime of 30 years , 2)
b b
federal tax rate of 35% , 3) state tax rate of 7% , 4) MACRS depreciation schedule, 5) no state or local
subsidies, 6) a working capital and debt service reserve account for six months of operating costs and
b
debt payments (earning an interest of 1.75%) , 7) a six month construction loan, with an interest rate of
b a
4% and a fee of 1% of the cost of the system , 8) a system size of 200 kW , 9) an inverter lifetime of 15
a
years , 10) a module tilt angle of 10 degrees, and an azimuth of 180 degrees, 11) debt with a term of 18
b
years , and 12) $1.1MM of upfront financial transaction costs for a $100MM TPO transaction of a pool of
d
commercial projects .
e
The financial assumptions in Table 7 assume a $100MM TPO transaction of a pool of commercial
projects.
f
The historical financial structure for a residential TPO system, assumed in 2010 from Feldman and
Bolinger 2016 does not assume a debt raise; however, the financial structure in 2016 from Feldman,
Lowder, and Schwabe does assume back-leveraged debt. To make these assumptions uniform, the “debt
interest rate” and “debt fraction” are taken from the utility-scale historical financial structure in Feldman
and Bolinger 2016 that uses back-leveraged debt.

32
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 As demonstrated in Table 9, in addition to a 65% reduction in installed cost from 2010 to 2017,
inverter replacement costs reduced 58%, O&M costs reduced 41%, annual degradation rates
reduced 25%, the equity discount rate reduced 23%, the debt interest rate reducd 13%, and the
debt fraction increased 17%.

Using these assumptions we calculated the LCOE, with and without the 30% federal investment
tax credit (ITC), in Phoenix, AZ, Kansas City, MO, and New York, NY, corresponding to
higher, medium, and lower resource areas in the United States and the locations used to calculate
LCOE in the SunShot Vision Study. The calculated values are summarized in Figure 23. 22

Figure 23. Levelized cost of energy for commercial PV systems, by region and with and
without ITC, 2010 –2017
As demonstrated in Figure 23, from 2010 to 2017 there was a 71% - 72% reduction in the
commercial PV system electricity cost benchmark (a 12% - 13% reduction was achieved from
2016 to 2017), bringing the unsubsidized LCOE between $0.09/kWh to $0.12/kWh ($0.06/kWh
to $0.08/kWh when including the federal ITC). This reduction is 89% toward achieving
SunShot’s 2020 commercial PV LCOE goal. 23

22
Because this analysis uses a more robust set of current and historical assumptions LCOE values may differ from
previously reported benchmarked values.
23
The SunShot 2020 target is adjusted from 2010 USD using the CPI. A Summary of these values can be found in
Appendix A and B. For LCOE Kansas City, MO, without ITC cases are $0.40/kWh in 2010 and $0.11/kWh in 2017
in 2017 USD from Appendix A and B. Thus, calculation is: (0.40 – 0.11)/(0.40 – 0.08) = 89%.

33
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 5 Utility-Scale PV Model
This section describes our utility-scale model’s structure, inputs, and assumptions (Section 5.1)
and output (Section 5.2).

5.1 Utility-Scale Model Structure, Inputs, and Assumptions

We model a 100-MW, 1,000-Vdc utility-scale system using 72-cell, multicrystalline 17.5%-
efficient modules from a Tier 1 supplier and three-phase central inverters. We model both fixed-
tilt and one-axis tracking on ground-mounted racking systems using driven-pile foundations. In
addition, we separate our cost estimate into EPC and project-development functions. Although
some firms engage in both activities in an integrated manner, we believe the distinction can help
separate and highlight the specific cost trends and drivers associated with each function.

Figure 24 presents a schematic of our utility-scale system cost model, and Table 10 details its
assumptions and inputs.

Figure 24. Utility-scale PV: model structure

34
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 10. Utility-Scale PV: Modeling Inputs and Assumptions

Category Modeled Value Description Sources

A large utility-scale system
System size >2 MW Model assumption
capacity
Module
17.5% Average module efficiency NREL (2017)
efficiency
Ex-factory gate (first buyer)
Module price $0.35/Wdc Bloomberg (2017), NREL (2017)
price, Tier 1 modules
$0.06/Wdc (fixed- Ex-factory gate prices (first
tilt) buyer) price, Tier 1 inverters Bloomberg (2017), NREL
Inverter price (2017), Bolinger and Seel
$0.06/Wdc (one- DC-to-AC ratio = 1.3 for both (2017)
axis tracker) fixed-tilt and one-axis tracker
$0.10–$0.21/Wdc
Structural for a 100-MW
Fixed-tilt racking or one-axis ASCE (2006), model
components system; varies by
tracking system assumptions, NREL (2017)
(racking) location and
system size
Conductors, conduit and
Electrical Varies by location fittings, transition boxes, Model assumptions, NREL
components and system size switchgear, panel boards, (2017), RSMeans (2016)
onsite transmission, etc.
8.67%–13% for
equipment and
material (except
EPC for transmission
Costs associated with EPC
overhead (% line costs); 23%–
SG&A, warehousing, shipping, NREL (2017)
of equipment 69% for labor
and logistics
costs) costs; varies by
system size, labor
activity, and
location
National benchmark applies an
Sales tax Varies by location average (by state) weighted by DSIRE (2017), RSMeans (2016)
2016 installed capacities
Electrician:
$19.37–$38.22
per hour Modeled labor rate assumes
Direct non-union and union labor and
Laborer: $12.64–
installation depends on state; national BLS (2017), NREL (2017)
$25.09 per hour
labor benchmark uses weighted
Varies by location average of state rates
and inverter
option
Workers compensation (state-
Burden rates weighted average), federal and
Total nationwide
(% of direct state unemployment RSMeans (2016)
average: 31.8%
labor) insurance, FICA, builders risk,
public liability

35
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Category Modeled Value Description Sources
$0.03–$0.09/Wdc For construction permits fee,
PII Varies by system interconnection, testing, and NREL (2017)
size and location commissioning
System size < 10 MW, use 0
miles for gen-tie line
Transmission $0.00–$0.02/Wdc
line System size > 200 MW, use 5 Model assumptions, NREL
Varies by system miles for gen-tie line (2017)
(gen-tie line) size
System size = 10–200 MW,
use linear interpolation

3%–12% Includes overhead expenses

such as payroll, facilities,
Developer Varies by system travel, legal fees, Model assumptions, NREL
overhead size (100 MW administrative, business (2017)
uses 3%; 5 MW development, finance, and
uses 12%) other corporate functions
Estimated as markup on EPC
Contingency 3% NREL (2017)
cost
5%–8% Applies a percentage margin to
Varies by system all costs including hardware,
Profit size (100 MW installation labor, EPC NREL (2017)
uses 5%; 5 MW overhead, developer overhead,
uses 8%) etc.

Figure 25 shows the percentage of U.S. utility-scale PV systems using tracking systems for
2007–2016. Although the data include one-axis and dual-axis tracking systems in the same
“tracking” category, there are many more one-axis trackers than dual-axis trackers (Bolinger and
Seel 2017). Cumulative tracking system installation reached 64% in 2016.

Figure 25. Percentage of U.S. utility-scale PV systems using tracking systems, 2007–2016
(Bolinger and Seel 2017)

36
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Although EPC contractors and developers tend to employ low-cost, non-union labor (based on
data from BLS 2017) for PV system construction when possible, union labor is sometimes
mandated. Construction trade unions may negotiate with the local jurisdiction and EPC
contractor/developer during the public review period of the permitting process. Figure 26 shows
2016 utility-scale PV capacity installed (GTM Research and SEIA 2017) and the proportion of
unionized labor in each state (BLS 2017). The unionized labor number represents the percentage
of employed workers in each state’s entire construction industry who are union members. In our
utility-scale model, both non-union and union labor rates are considered (Figure 27).

Figure 26. Utility-scale PV: 2016 capacity installed and percentage of unionized labor by state
(BLS 2017; GTM Research and SEIA 2017)

5.2 Utility-Scale Model Output

Figure 27 presents the regional EPC benchmark from our utility-scale model, and Figure 28
presents the U.S. national benchmark (EPC + developer) for fixed-tilt and one-axis tracker
systems, using non-union labor. In Figure 28, note the following:

1. The national benchmark applies an average weighted by 2016 installed capacities.

2. Non-union labor is used.
3. Economies of scale—driven by BOS, labor, related markups, and development cost—are
demonstrated.
As in the commercial PV sector, the 7% net profit in Q1 2017 is higher than the 2% in Q1 2016,
because the rapid module price reduction in 2016 enabled EPC firms and developers to retain a
higher profit and still keep a competitive project cost bid.

37
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 $0.70 $0.80 $0.90 $1.00 $1.10 $1.20 $1.30 $1.40
Arkansas
Oklahoma
Alabama Modeled EPC Costs
North Carolina For 100 MW Systems
South Dakota
Georgia (2017$ / Wdc)
Texas
Montana
Louisiana
(1) Fixed-tilt & Non-Union Labor
Wyoming
Arizona (2) One-Axis Tracker & Non-Union Labor
Oregon
Utah
New Mexico (3) One-Axis Tracker & Union Labor
Tennessee
Virginia
Colorado
Delaware
Nebraska
South Carolina
Kansas
New Hampshire
Idaho
Mississippi
Vermont
Kentucky
Iowa
Puerto Rico
Maryland
Florida
West Virginia
Ohio
Maine
Missouri
Indiana
Michigan
North Dakota
Pennsylvania
Nevada
DC
Washington
Wisconsin
California
Rhode Island
New York
Illinois
Minnesota
New Jersey
Hawaii
Connecticut
Massachusetts
Alaska

Figure 27. Q1 2017 benchmark by location: 100-MW utility-scale PV systems, EPC only
24
(2017 USD/Wdc)

24
The fixed-tilt, non-union cost is always lowest, followed by the one-axis tracker, non-union cost and the one-axis
tracker, union cost. Thus the bars are additive: the fixed-tilt, non-union cost is represented by the dark green bar
alone; the one-axis tracker, non-union cost is the sum of the dark green and medium green bars; and the one-axis
tracker, union cost is the sum of all three bars.

38
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 25
Figure 28. Q1 2017 U.S. benchmark: utility-scale PV total cost (EPC + developer), 2017 USD/Wdc

25
Although four different system sizes are shown in this figure, the actual national average system size in 2015 was
29 MW for fixed-tilt systems and 37 MW for one-axis tracker systems. Our model estimates $1.17/W for 29-MW
fixed-tilt systems and $1.25/W for 37-MW one-axis tracker systems.

39
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 5.3 Utility-Scale PV Price Benchmark Historical Trends
Figure 29 summarizes the reduction in utility-scale PV system cost benchmarks between 2010
and 2017. 26

Figure 29. NREL utility-scale PV system cost benchmark summary (inflation adjusted), 2010–2017
As demonstrated in Figure 29, from 2010 to 2017 there was a 77% reduction in the utility-scale
(fixed-tilt) PV system cost benchmark, and an 80% reduction in the utility-scale (one-axis) PV
system cost benchmark. Approximately 71% and 64% of that reduction can be attributed to total
hardware costs (for fixed-tilt and one-axis systems respectively), as module prices dropped 86%
over that time period. An additional 10% / 11% can be attributed to labor, which dropped 74% /
78% over that time period, with the final 19% / 25% attribitubal to other soft costs, including PII,
sales tax, overhead, and net profit (for fixed-tilt and one-axis systems respectively).

Looking at this past year, from 2016 to 2017 there was a 29% reduction in the utility-scale
(fixed-tilt) PV system cost benchmark, and an 28% reduction in the utility-scale (one-axis) PV
system cost benchmark. The majority of that reduction can be attributed to the 46% reduction in
module factory gate price, and a 45% / 41% reduction in inverter factory gate price. 27

26
Each year’s PV system cost benchmark corresponds to the NREL benchmark calculted in Q4 of the previous year
or Q1 of the current year (e.g. 2010 = Q4 2009; 2017 = Q1 2017).
27
One-axis and fixed-tilt PV systems have different reductions in inverter factory gate price due to differing ILRs
in 2016.

40
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 5.4 Utility-Scale PV Levelized Cost of Energy Historical Trends
While LCOE is not a perfect metric to measure the competiveness of PV within the energy
marketplace, it incorporates many other PV metrics important to the energy costs beyond upfront
installation costs. These benchmarks are summarized over time in Table 11 (next page), from Q4
2009 to Q1 2017 (aSunShot Vision Study 2010, bOn the Path to SunShot: The Role of
Advancements in Solar Photovoltaic Efficiency, Reliability, and Costs; cOn the Path to SunShot:
Emerging Opportunities and Challenges in Financing Solar (Feldman and Bolinger 2016);
d
Terms, Trends, and Insights PV Project Finance in the United States (Feldman, Lowder and
Schwabe 2016), eU.S. Solar Photovoltaic System Cost Benchmark: Q1 2016). 28

As demonstrated in Table 11, in addition to a 80% reduction in installed cost of utility-scale

(one-axis) systems from 2010 to 2017, inverter replacement costs reduced 68%, O&M costs
reduced 17%, annual degradation rates reduced 25%, the equity discount rate reduced 14%, the
debt interest rate reducd 18%, and the debt fraction increased 17%.

Using these assumptions we calculated the LCOE, with and without the 30% federal investment
tax credit (ITC), in Phoenix, AZ, Kansas City, MO, and New York, NY, corresponding to
higher, medium, and lower resource areas in the United States and the locations used to calculate
LCOE in the SunShot Vision Study. The calculated values are summarized in Figure 30. 29

28
In instances in which LCOE assumptions were not found from the selected literature in a given year, straight-line
changes were assumed between any two values.
29
Because this analysis uses a more robust set of current and historical assumptions LCOE values may differ from
previously reported benchmarked values.

41
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 11. One-Axis Tracker and Fixed-Tilt Utility-Scale PV LCOE Assumptions, 2010–2017

2017 USD per Watt DC 2010 2011 2012 2013 2014 2015 2016 2017
One-Axis Tracker
Installed cost $5.44 $4.59 $3.15 $2.39 $2.15 $1.97 $1.54 $1.11
a b
Annual degradation (%) 1.00% 0.95% 0.90% 0.85% 0.80% 0.75% 0.75% 0.75%
a b e
Inverter replacement price ($/W) $0.19 $0.17 $0.15 $0.14 $0.12 $0.10 $0.08 $0.06
a b
O&M expenses ($/kw-yr) $22.2 $21.5 $20.7 $20.0 $19.2 $18.5 $18.5 $18.5
a b
Pre-inverter derate (%) 90.5% 90.50% 90.50% 90.50% 90.50% 90.5% 90.5% 90.5%
a b
Inverter efficiency (%) 96.0% 96.40% 96.80% 97.20% 97.60% 98.0% 98.0% 98.0%
a e
Inverter loading ratio 1.10 1.12 1.13 1.15 1.17 1.18 1.20 1.30
c d
Equity discount rate (real) 7.4% 7.2% 7.0% 6.9% 6.7% 6.5% 6.3% 6.3%
a
Inflation rate 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5%
c d
Debt interest rate 5.5% 5.3% 5.2% 5.0% 4.8% 4.7% 4.5% 4.5%
c d
Debt fraction 34.2% 35.2% 36.1% 37.1% 38.1% 39.0% 40.0% 40.0%
Fixed-Tilt
Installed cost $4.57 $3.91 $2.66 $2.04 $1.89 $1.82 $1.45 $1.03
a b
Annual degradation (%) 1.00% 0.95% 0.90% 0.85% 0.80% 0.75% 0.75% 0.75%
a b e
Inverter replacement price ($/W) $0.19 $0.17 $0.15 $0.14 $0.12 $0.10 $0.08 $0.06
a b
O&M expenses ($/kw-yr) $22.2 $20.9 $19.5 $18.1 $16.8 $15.4 $15.4 $15.4
a b
Pre-inverter derate (%) 90.5% 90.50% 90.50% 90.50% 90.50% 90.5% 90.5% 90.5%
a b
Inverter efficiency (%) 96.0% 96.40% 96.80% 97.20% 97.60% 98.0% 98.0% 98.0%
a e
Inverter loading ratio 1.10 1.15 1.2 1.25 1.3 1.35 1.40 1.3
c d
Equity discount rate (real) 7.4% 7.2% 7.0% 6.9% 6.7% 6.5% 6.3% 6.3%
a
Inflation rate 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5% 2.5%
c d
Debt interest rate 5.5% 5.3% 5.2% 5.0% 4.8% 4.7% 4.5% 4.5%
c d
Debt fraction 34.2% 35.2% 36.1% 37.1% 38.1% 39.0% 40.0% 40.0%
a b b
Other important assumptions: utility-scale PV system LCOE assume a 1) system lifetime of 30 years , 2) federal tax rate of 35% , 3) state tax rate of 7% , 4) MACRS depreciation
b
schedule, 5) no state or local subsidies, 6) a working capital and debt service reserve account for six months of operating costs and debt payments (earning an interest of 1.75%) ,
b b a
7) a six month construction loan, with an interest rate of 4% and a fee of 1% of the cost of the system , 8) a system size of 100 MW , 9) an inverter lifetime of 15 years , 10) debt
b d
with a term of 18 years , and 11) $1.1MM of upfront financial transaction costs .

42
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 30. Levelized cost of energy for utility-scale PV systems, by region and with
and without ITC, 2010–2017
We use the fixed-tilt systems for LCOE benchmarks from 2010 to 2015 and then switch to one-
axis tracking systems from 2016 to 2017 to reflect the market share change in Figure 31. All
detailed LCOE values can be found in Appendix A and B.
As demonstrated in Figure 30, from 2010 to 2017 there was a 78%–79% reduction in the utility-
scale PV system electricity cost benchmark (a 20%– 23% reduction was achieved from 2016 to
2017), bringing the unsubsidized LCOE between $0.04/kWh to $0.06/kWh ($0.03/kWh to
$0.04/kWh when including the federal ITC). This reduction signifies the achievement of
SunShot’s 2020 utility-scale PV goal. 30

30
The 2020 utility-scale goal is not adjusted for inflation as wholesale prices have been relatively flat, and in some
cases gone down, from 2010-2017. A Summary of these values can be found in Appendix A and B.

43
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 6 Model Applications
This section includes three additional applications of our cost modeling: system cost reduction
from economies of scale (Section 6.1), module efficiency impacts (Section 6.2), and regional
LCOE (Section 6.3). The granularity of our bottom-up models enables us to determine the
changes in particular cost drivers over time. Accordingly, the models can be used to predict
future system cost-reduction opportunities based on particular market trends and technologies.

6.1 System Cost Reduction from Economies of Scale

Figure 31 demonstrates the cost savings from increased system size. Scaling up the system size
from 10 MW to 100 MW reduces related costs in several ways: per-watt BOS costs because of
bulk purchasing, labor costs that benefit from learning-related improvements for larger systems,
and EPC overhead and developer costs that are spread over more installed capacity. Note that
non-union labor is used in this figure.

Figure 32. Model application: U.S. utility-scale one-axis tracking PV system cost reduction
from economies of scale (2017 USD/Wdc)

6.2 Module Efficiency Impacts

Our system cost models can also assess the economic benefits of high module efficiency.
Because higher module efficiency reduces the number of modules required to reach a certain
system size, the related racking or mounting hardware, foundation, BOS, EPC/developer
overhead, and labor hours are reduced accordingly. Figure 32 presents the relationship between
module efficiency and installed cost (with module prices held equal for any given efficiency) and
demonstrates the cost-reduction potential due to high module efficiency. Note that a fixed-tilt
system is used in the utility-scale curve and a string inverter is used in the residential curve.

44
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 33. Modeled impacts of module efficiency on total system costs, 2017

45
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 6.3 Regional LCOE
To estimate regional LCOEs across the United States, we combine modeled regional installed
cost with localized solar irradiance and weather data, a PV performance model, and a pro forma
financial analysis that models the revenue, operating expenses, taxes, incentives, debt structures,
and cash flows for a representative PV system. We use NREL’s System Advisor Model (SAM),
a performance and financial model, 31 to estimate location-specific hourly energy output over the
PV system’s lifetime and subsequently calculate the resulting real LCOEs (considering inflation)
for each location. Figure 33 presents real LCOEs for a 100-MW utility-scale PV system with
fixed tilt or one-axis tracking based on regional labor and material costs, wind speeds, snow
loading, solar irradiance, weather data, and sales tax. 32 We assume the following:

• ITC = 0%, Real discount rate = 6.3%, IRR target = 6.46%, Inflation = Price escalator =
2.5%, Analysis period = 30-Yr, Degradation rate = 0.75% per year. System size = 100
MW utility-scale PV, Project debt = 40%, Debt interest rate = 4.5%.
• Fixed-tilt: DC-to-AC ratio = 1.3 and Fixed O&M cost = $15/kW per year. One-axis
tracker: DC-to-AC ratio = 1.3 and Fixed O&M cost = $18.5/kW per year.

31
See https://sam.nrel.gov/.
32
The assumptions in this LCOE exercise are the same from those in Section 5.

46
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 ¢/kWh
Real LCOE
(Fixed-tilt)

Real LCOE
(One-axis tracker)

2017 USD Fixed-Tilt One-Axis Tracker One-Axis Tracker vs. Fixed-Tilt

Total Installed Nominal LCOE Real LCOE Total Installed Nominal LCOE Real LCOE Installed Costs Nominal LCOE Real LCOE
State Location Costs ($/W) (cent per kWh) (cent per kWh) Costs ($/W) (cent per kWh) (cent per kWh) Premium (%) Change (%) Change (%)
CA Bakersfield 1.09 7.26 5.68 1.18 6.44 5.04 8.26% -11.29% -11.27%
CA Imperial 1.09 6.64 5.19 1.18 5.76 4.50 8.26% -13.25% -13.29%
AZ Prescott 0.98 6.20 4.85 1.06 5.47 4.27 8.16% -11.77% -11.96%
AZ Tucson 0.98 6.01 4.70 1.06 5.29 4.14 8.16% -11.98% -11.91%
NV Las Vegas 1.05 6.33 4.95 1.13 5.54 4.33 7.62% -12.48% -12.53%
NM Albuquerque 0.99 6.05 4.73 1.06 5.39 4.21 7.07% -10.91% -10.99%
CO Alamosa 0.99 6.05 4.73 1.07 5.33 4.16 8.08% -11.90% -12.05%
NC Jacksonville 0.96 7.25 5.67 1.03 6.56 5.13 7.29% -9.52% -9.52%
TX San Antonio 0.97 7.11 5.56 1.04 6.55 5.12 7.22% -7.88% -7.91%
NJ Newark 1.13 9.15 7.16 1.22 8.59 6.71 7.96% -6.12% -6.28%
FL Orlando 1.02 8.47 6.63 1.09 7.51 5.87 6.86% -11.33% -11.46%
HI Kona 1.14 8.08 6.32 1.22 7.41 5.79 7.02% -8.29% -8.39%

Figure 34. Modeled real LCOE (¢/kWh), ITC = 0%, for a 100-MWdc utility-scale PV system with
33
fixed-tilt and one-axis tracking in 2017

33
The U.S. Department of Energy’s SunShot Initiative uses Kansas City’s insolation as the national average
insolation to calculate LCOE (Woodhouse et al. 2016).

47
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 7 Conclusions
Based on our bottom-up modeling, the Q1 2017 PV cost benchmarks are $2.80/Wdc (or
$3.22/Wac) for residential systems, $1.85/Wdc (or $2.13/Wac) for commercial systems,
$1.03/Wdc (or $1.34/Wac) for fixed-tilt utility-scale systems, and $1.11/Wdc (or $1.44/Wac) for
one-axis-tracking utility-scale systems. Overall, modeled PV installed costs continued to decline
in Q1 2017 for all three sectors.

Figure 34 puts our Q1 2017 benchmark results in context with the results of previous NREL
benchmarking analyses. When comparing the results across this period, note the following:

1. Values are inflation adjusted using the U.S. Bureau of Labor Statistics’ Consumer Price
Index. Thus, historical values from our models are adjusted and presented as real USD
instead of as nominal USD.
2. Cost categories are aggregated for comparison purposes. “Soft Costs – Others” represents
PII, land acquisition, sales tax, and EPC/developer overhead and profit. 34
3. The “Utility-Scale PV, One-Axis Tracker (100 MW)” consists of our previous bottom-up
results (2010 and 2013–2016) and interpolation estimates for 2009 and 2011–2012.
4. The comparison of Q1 2016 and Q1 2017 is presented in Table 12.
The inflation-adjusted system cost differences between Q1 2016 and Q1 2017 are $0.18/Wdc
(residential), $0.32/Wdc (commercial), and $0.42/Wdc (fixed-tilt utility-scale). Table 12 shows
the benchmarked values for all three sectors and drivers of cost decrease and increase.

As Figure 34 shows, hardware costs—and module prices in particular—declined substantially in

Q1 2017 owing to an imbalance in global module supply and demand. This has increased the
importance of non-hardware, or “soft,” costs. 35 Figure 35 shows the growing contribution from
soft costs. 36 Soft costs and hardware costs also interact with each other. For instance, module
efficiency improvements have reduced the number of modules required to construct a system of
a given size, thus reducing hardware costs. This trend has also reduced soft costs from direct
labor and related installation overhead.

Also, our bottom-up system cost models enable us to investigate regional variations, system
configurations (such as MLPE vs. non-MLPE, fixed-tilt vs. one-axis tracker, and small vs. large
system size). And, business structures (such as installer vs. integrator, and EPC vs. developer)
are considered. Different scenarios result in different costs, so consistent comparisons can only
be made when cost scenarios are aligned.

34
System cost categories in this report differ from previously published material, beyond inflation adjustments, to
delineate profit from overhead for installers and integrators. Also, profit is added to the Q1 2015 commercial
benchmark price; thus it is $0.06/W higher than in the 2015 publication ($0.05/W profit, $0.01/W inflation).
35
Soft cost = total cost - hardware (module, inverter, structural, and electrical BOS) cost.
36
An increasing soft cost proportion in Figure 35 indicates soft costs declined more slowly than did hardware costs;
it does not indicate soft costs increased on an absolute basis.

48
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Finally, the reduction in installed cost, along with improvements in operation, system design, and
technology have resulted in significant reduction in the cost of electricity, as shown in Figure 36.
U.S. residential and commercial PV systems are 86% and 89% toward achieving SunShot’s 2020
electricity price targets, and U.S. utility-scale PV systems have achieved their 2020 SunShot
target three years early.

Figure 35. NREL PV system cost benchmark summary (inflation adjusted), 2010–2017

49
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 12. Comparison of Q1 2016 and Q1 2017 PV System Cost Benchmarks

Sector Residential PV Commercial PV Utility-Scale PV, Fixed-Tilt

Q1 2016 Benchmarks in
$2.93 $2.13 $1.42
2016 USD/Wdc

Q1 2016 Benchmarks in
$2.98 $2.17 $1.45
2017 USD/Wdc

Q1 2017 Benchmarks in
$2.80 $1.85 $1.03
2017 USD/Wdc

• Lower module • Lower module • Lower module price

price price • Lower inverter price
• Lower inverter • Lower inverter • Higher module
price price efficiency
• Higher module • Higher module
efficiency efficiency
• Lower electrical • Smaller
BOS commodity developer team
Drivers of Cost Decrease price
• Higher small
installer market
share
• Lower sales &
marketing costs
• Lower overhead
(general &
administrative)

• Higher labor • Higher labor • Higher labor wages

wages wages • Higher net profit
• Higher advanced • Higher PII costs
inverter adoption • Higher net profit
Drivers of Cost Increase • More BOS
components for
rapid shutdown
• Higher supply-
chain costs

50
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Figure 36. Modeled trend of soft cost as a proportion of total cost by sector, 2010–2017

Figure 37. NREL PV LCOE benchmark summary (inflation adjusted), 2010–2017

51
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 References
ASCE (American Society of Civil Engineers). 2006. Minimum Design Loads for Buildings and
Other Structures (7-05). Reston, VA: American Society of Civil Engineers.
Barbose, G., and N. Darghouth. 2016. Tracking the Sun IX: The Installed Price of Residential
and Non-Residential Photovoltaic Systems in the United States. Berkeley, CA: Lawrence
Berkeley National Laboratory. https://emp.lbl.gov/publications/tracking-sun-ix-installed-price.
Bloomberg. 2017. Bloomberg Professional service. Accessed May 20, 2017.
BLS (U.S. Bureau of Labor Statistics). 2017. “Occupational Employment Statistics.” Accessed
May 8, 2017. http://www.bls.gov/oes/tables.htm.
Bolinger, M., and J. Seel. 2017 (forthcoming). Utility-Scale Solar 2016: An Empirical Analysis
of Project Cost, Performance, and Pricing Trends in the United States. Berkeley, CA: Lawrence
Berkeley National Laboratory.
———. 2016. Utility-Scale Solar 2015: An Empirical Analysis of Project Cost, Performance,
and Pricing Trends in the United States. Berkeley, CA: Lawrence Berkeley National Laboratory.
https://emp.lbl.gov/publications/utility-scale-solar-2015-empirical.
Case, T. 2012. “US Cost of Doing Business: Costs Fall in 2010.” Moody’s Analytics Regional
Financial Review, September 2012.
Chung, D., C. Davidson, R. Fu, K. Ardani, and R. Margolis. 2015. U.S. Photovoltaic Prices and
Cost Breakdowns: Q1 2015 Benchmarks for Residential, Commercial, and Utility-Scale Systems.
NREL/TP-6A20-64746. Golden, CO: National Renewable Energy Laboratory.
http://www.nrel.gov/docs/fy15osti/64746.pdf.
CSI (California Solar Initiative). 2017. “CSI Working Data Set.” Accessed May 10, 2017.
http://www.californiadgstats.ca.gov/downloads.
DSIRE (Database of State Incentives for Renewables & Efficiency). 2017. Accessed May 10,
2017. http://www.dsireusa.org/.
Enphase. 2017. Enphase quarterly presentations. Accessed May 20, 2017.
http://investor.enphase.com/events.cfm?Year=2017.
Feldman, D., G. Barbose, R. Margolis, M. Bolinger, D. Chung, R. Fu, J. Seel, C. Davidson,
N. Darghouth, and R. Wiser. 2015. Photovoltaic System Pricing Trends: Historical, Recent, and
Near-Term Projections. NREL/PR-6A20-64898. Golden, CO: National Renewable Energy
Laboratory. http://www.nrel.gov/docs/fy15osti/64898.pdf.
Feldman, D., B. Friedman, and R. Margolis. 2013. Financing, Overhead, and Profit: An In-
Depth Discussion of Costs Associated with Third-Party Financing of Residential and
Commercial Photovoltaic Systems. NREL/TP-6A20-60401. Golden, CO: National Renewable
Energy Laboratory. http://www.nrel.gov/docs/fy14osti/60401.pdf.
Feldman, D., Lowder, T. Lowder, and P. Schwabe. 2016. Terms, Trends, and Insights PV
Project Finance in the United States, 2016. NREL/BR-6A20-66991. Golden, CO: National
Renewable Energy Laboratory. http://www.nrel.gov/docs/fy16osti/66991.pdf.

52
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Fu, R., D. Chung, T. Lowder, D. Feldman, K. Ardani, and R. Margolis. 2016. U.S. Solar
Photovoltaic System Cost Benchmark: Q1 2016. NREL/TP-6A20-66532. Golden, CO: National
Renewable Energy Laboratory. http://www.nrel.gov/docs/fy16osti/66532.pdf.
Fu, R., T.L. James, D. Chung, D. Gagne, A. Lopez, and A. Dobos. 2015. “Economic
Competitiveness of U.S. Utility-Scale Photovoltaic Systems in 2015: Regional Cost Modeling of
Installed Cost ($/W) and LCOE ($/kWh).” Presented at the IEEE 42nd Photovoltaic Specialist
Conference, New Orleans, LA.
http://www.nrel.gov/analysis/pdfs/Economic_Competitiveness_of_US_Utility-
Scale_Photovoltaics_System.pdf.
Go Solar CA (Go Solar California). 2017. “Currently Interconnected Data Set.” Accessed May
10, 2017. http://www.californiadgstats.ca.gov/downloads.
GTM Research. 2017. Q1 2017 Solar Executive Briefing. Boston: Greentech Media.
GTM Research and SEIA (Solar Energy Industries Association). 2017. U.S. Solar Market Insight
Report, Q1 2017. Washington, D.C.: Solar Energy Industries Association.
NREL. 2017. NREL dialogues and interviews with solar industry collaborators. Golden, CO:
National Renewable Energy Laboratory.
PVinsights. 2017. PVinsights database. Accessed May 20, 2017.
RSMeans, ed. 2016. RSMeans Building Construction Cost Data 2015, 73rd annual edition.
Norwell, MA: RSMeans.
SolarEdge. 2017. SolarEdge quarterly presentations. Accessed May 20, 2017.
http://investors.solaredge.com/phoenix.zhtml?c=253935&p=irol-news&nyo=0.
Sunrun. 2017. Sunrun quarterly presentations. Accessed May 20, 2017.
http://investors.sunrun.com/phoenix.zhtml?c=254007&p=irol-calendar.
Vivint Solar. 2017. Vivint Solar quarterly presentations. Accessed May 20, 2017.
http://investors.vivintsolar.com/company/investors/events-and-
presentations/presentations/default.aspx.
Vote Solar. 2015. “Project Permit.” Accessed May 8, 2017.
http://projectpermit.org/2013/02/06/best-practices/.
Vote Solar and IREC (Interstate Renewable Energy Council). 2013. Project Permit: Best
Practices in Residential Solar Permitting. San Francisco: Vote Solar Initiative.
Woodhouse, M., R. Jones-Albertus, D. Feldman, R. Fu, K. Horowitz, D. Chung, D. Jordan, and
S. Kurtz. 2016. On the Path to SunShot: The Role of Advancements in Solar Photovoltaic
Efficiency, Reliability, and Costs. NREL/TP-6A20-65872. Golden, CO: National Renewable
Energy Laboratory. http://www.nrel.gov/docs/fy16osti/65872.pdf.
XE Currency Charts. 2017. Historical currency conversion. Accessed May 2, 2017.
http://www.xe.com/currencycharts/.

53
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Appendix A. Historical PV System Benchmarks in
2010 USD
Table 13. NREL Residential PV Benchmark Summary (Inflation Adjusted), 2010–2017

2010 USD per Watt DC 2010 2011 2012 2013 2014 2015 2016 2017
Module $2.26 $1.89 $0.98 $0.68 $0.65 $0.63 $0.57 $0.31
Inverter $0.41 $0.60 $0.40 $0.38 $0.28 $0.26 $0.19 $0.17
Hardware BOS - Structural and $0.49 $0.45 $0.42 $0.46 $0.42 $0.30 $0.33 $0.31
Electrical Components
Soft Costs - Install Labor $0.99 $0.62 $0.59 $0.73 $0.29 $0.30 $0.26 $0.27
Soft Costs - Others (PII, Sales $2.22 $2.01 $1.54 $1.20 $1.37 $1.31 $1.26 $1.40
Tax, Overhead, and Net Profit)
Total $6.36 $5.58 $3.94 $3.44 $3.02 $2.80 $2.61 $2.45
Total Inverter Replacement Price $0.37 $0.32 $0.28 $0.23 $0.18 $0.14 $0.13 $0.12
($/W)
O&M Expenses ($/kW-yr) $33 $30 $27 $24 $21 $18 $18 $18
LCOE Phoenix, AZ, no ITC $0.38 $0.32 $0.22 $0.19 $0.15 $0.13 $0.12 $0.12
LCOE Kansas City, MO, no ITC $0.46 $0.39 $0.27 $0.23 $0.19 $0.16 $0.15 $0.14
LCOE New York, NY, no ITC $0.49 $0.42 $0.29 $0.24 $0.20 $0.17 $0.16 $0.15
LCOE Phoenix, AZ, ITC $0.24 $0.20 $0.14 $0.12 $0.10 $0.09 $0.08 $0.07
LCOE Kansas City, MO, ITC $0.30 $0.25 $0.18 $0.15 $0.12 $0.10 $0.09 $0.09
LCOE New York, NY, ITC $0.32 $0.27 $0.19 $0.16 $0.13 $0.11 $0.10 $0.10

54
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 13. NREL Commercial PV Benchmark Summary (Inflation Adjusted), 2010–2017

2010 USD per Watt DC 2010 2011 2012 2013 2014 2015 2016 2017
Module $2.23 $1.89 $0.98 $0.59 $0.64 $0.62 $0.57 $0.31
Inverter $0.32 $0.37 $0.27 $0.24 $0.15 $0.12 $0.12 $0.09
Hardware BOS - Structural and
Electrical Components $0.63 $0.64 $0.60 $0.59 $0.38 $0.33 $0.29 $0.26
Soft Costs - Install Labor $0.28 $0.28 $0.27 $0.26 $0.19 $0.17 $0.17 $0.15
Soft Costs - Others (PII, Sales
Tax, Overhead, and Net Profit) $1.25 $1.18 $0.88 $0.75 $1.06 $0.76 $0.76 $0.81
Total $4.71 $4.36 $3.00 $2.44 $2.42 $1.99 $1.90 $1.62
Total Inverter Replacement
Price ($/W) $0.22 $0.19 $0.17 $0.15 $0.13 $0.11 $0.10 $0.09
O&M Expenses ($/kW-yr) $24 $22 $20 $18 $16 $14 $14 $14
LCOE Phoenix, AZ, no ITC $0.29 $0.26 $0.17 $0.14 $0.13 $0.10 $0.09 $0.08
LCOE Kansas City, MO, no ITC $0.36 $0.32 $0.22 $0.17 $0.16 $0.12 $0.12 $0.10
LCOE New York, NY, no ITC $0.38 $0.34 $0.23 $0.18 $0.17 $0.13 $0.12 $0.11
LCOE Phoenix, AZ, ITC $0.18 $0.16 $0.11 $0.09 $0.08 $0.06 $0.06 $0.05
LCOE Kansas City, MO, ITC $0.23 $0.20 $0.14 $0.11 $0.10 $0.08 $0.07 $0.07
LCOE New York, NY, ITC $0.24 $0.21 $0.15 $0.12 $0.11 $0.08 $0.08 $0.07

55
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
 Table 14. NREL Utility-Scale PV Benchmark Summary (Inflation Adjusted), 2010–2017

2010 USD per 2010 2011 2012 2013 2014 2015 2016 2017
Watt DC
Module $2.12 $1.89 $0.98 $0.59 $0.60 $0.59 $0.57 $0.31
Inverter $0.24 $0.28 $0.24 $0.16 $0.11 $0.10 $0.10 $0.05
Hardware BOS - $0.66 $0.58 $0.50 $0.43 $0.37 $0.34 $0.22 $0.22
Structural and
Electrical
Components
Soft Costs - $0.54 $0.48 $0.45 $0.44 $0.21 $0.18 $0.14 $0.12
Install Labor
Soft Costs - $1.22 $0.81 $0.59 $0.48 $0.59 $0.52 $0.31 $0.27
Others (PII, Land
Acquisition, Sales
Tax, Overhead,
and Net Profit)
Total $4.78 $4.03 $2.77 $2.10 $1.88 $1.73 $1.35 $0.97
Total Inverter $0.17 $0.15 $0.14 $0.12 $0.11 $0.09 $0.07 $0.05
Replacement
Price ($/W)
O&M Expenses $20 $19 $19 $18 $17 $17 $17 $17
($/kW-yr)
LCOE Phoenix, $0.19 $0.16 $0.11 $0.08 $0.07 $0.06 $0.05 $0.04
AZ, no ITC
LCOE Kansas $0.24 $0.20 $0.14 $0.10 $0.09 $0.08 $0.06 $0.05
City, MO, no ITC
LCOE New York, $0.26 $0.22 $0.15 $0.12 $0.10 $0.09 $0.07 $0.06
NY, no ITC
LCOE Phoenix, $0.12 $0.10 $0.07 $0.05 $0.05 $0.04 $0.03 $0.03
AZ, ITC
LCOE Kansas $0.15 $0.12 $0.09 $0.07 $0.06 $0.05 $0.04 $0.03
City, MO, ITC
LCOE New York, $0.17 $0.14 $0.10 $0.08 $0.07 $0.06 $0.05 $0.04
NY, ITC

56
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
Appendix B. PV System LCOE Benchmarks in 2017 and 2010 USD
Table 16. NREL LCOE Summary (2017 cents/kWh)

37
2020 2030
Reporting Year 2010 2011 2012 2013 2014 2015 2016 2017 Goal Goal
Benchmark Date Q4 2009 Q4 2010 Q4 2011 Q4 2012 Q4 2013 Q1 2015 Q1 2016 Q1 2017
Residential
Phoenix, AZ, no ITC 42.1 35.7 24.9 20.7 17.3 15.0 13.6 12.9
Kansas City, MO, no ITC 51.8 43.6 30.4 25.3 21.1 18.3 16.7 15.7 10.0 5.0
New York, NY, no ITC 55.2 46.5 32.4 26.9 22.4 19.5 17.7 16.7
Phoenix, AZ, ITC 26.9 22.8 16.1 13.4 11.1 9.5 8.7 8.2
Kansas City, MO, ITC 33.1 27.9 19.7 16.3 13.5 11.6 10.6 10.0
New York, NY, ITC 35.3 29.7 21.0 17.4 14.4 12.3 11.3 10.7
Commercial
Phoenix, AZ, no ITC 32.3 28.6 19.5 15.4 14.4 11.2 10.5 9.2
Kansas City, MO, no ITC 40.0 35.3 24.1 19.0 17.8 13.9 13.0 11.3 7.8 4.0
New York, NY, no ITC 42.4 37.5 25.6 20.2 18.9 14.8 13.8 12.0
Phoenix, AZ, ITC 20.4 18.0 12.5 9.9 9.2 7.1 6.7 5.9
Kansas City, MO, ITC 25.2 22.2 15.4 12.3 11.4 8.9 8.3 7.3
New York, NY, ITC 26.8 23.6 16.4 13.0 12.0 9.4 8.8 7.7
Utility-scale (one-axis tracking)
Phoenix, AZ, no ITC 21.2 17.5 12.1 9.2 8.1 7.2 5.7 4.4
Kansas City, MO, no ITC 26.8 22.1 15.3 11.7 10.2 9.1 7.2 5.6 6.0 3.0
New York, NY, no ITC 29.5 24.3 16.8 12.9 11.3 10.0 7.9 6.1
Phoenix, AZ, ITC 13.4 11.0 7.8 6.0 5.3 4.7 3.8 3.0
Kansas City, MO, ITC 16.9 13.9 9.8 7.6 6.7 5.9 4.8 3.8
New York, NY, ITC 18.6 15.4 10.8 8.4 7.4 6.5 5.3 4.2

37
2020 Residential and commercial SunShot goals are adjusted for inflation using the Consumer Price Index; the 2020 utility-scale goal was left unchanged as wholesale prices
have been relatively flat, and in some cases gone down, from 2010-2017.
57
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
Utility-scale (fixed-tilt)
Phoenix, AZ, no ITC 22.6 18.9 13.0 10.1 9.0 8.4 6.8 5.0
Kansas City, MO, no ITC 27.7 23.1 15.9 12.3 11.0 10.2 8.3 6.1
New York, NY, no ITC 29.6 24.7 17.0 13.2 11.8 10.9 8.8 6.6
Phoenix, AZ, ITC 14.4 12.0 8.5 6.6 5.9 5.4 4.5 3.4
Kansas City, MO, ITC 17.6 14.7 10.4 8.1 7.3 6.7 5.4 4.2
New York, NY, ITC 18.9 15.8 11.1 8.7 7.8 7.1 5.8 4.4
Residential
Phoenix, AZ, no ITC 37.8 32.0 22.3 18.5 15.5 13.4 12.2 11.5
Kansas City, MO, no ITC 46.4 39.1 27.3 22.7 18.9 16.4 14.9 14.1 9.0 5.0
New York, NY, no ITC 49.5 41.6 29.0 24.1 20.1 17.4 15.9 15.0
Phoenix, AZ, ITC 24.1 20.4 14.5 12.0 9.9 8.5 7.8 7.4
Kansas City, MO, ITC 29.7 25.0 17.7 14.6 12.1 10.4 9.5 9.0
New York, NY, ITC 31.6 26.6 18.8 15.6 12.9 11.1 10.1 9.6
Commercial
Phoenix, AZ, no ITC 29.0 25.6 17.5 13.8 12.9 10.1 9.4 8.2
Kansas City, MO, no ITC 35.8 31.7 21.6 17.0 16.0 12.5 11.7 10.1 7.0 4.0
New York, NY, no ITC 38.0 33.6 22.9 18.1 16.9 13.3 12.4 10.7
Phoenix, AZ, ITC 18.3 16.1 11.2 8.9 8.2 6.4 6.0 5.3
Kansas City, MO, ITC 22.6 19.9 13.8 11.0 10.2 8.0 7.4 6.5
New York, NY, ITC 24.0 21.1 14.7 11.6 10.8 8.4 7.9 6.9
Utility-scale (one-axis
38
tracking)
Phoenix, AZ, no ITC 19.0 15.6 10.8 8.3 7.2 6.4 5.1 3.9
Kansas City, MO, no ITC 24.0 19.8 13.7 10.5 9.2 8.1 6.4 5.0 6.0 3.0
New York, NY, no ITC 26.4 21.8 15.1 11.5 10.1 9.0 7.1 5.5
Phoenix, AZ, ITC 12.0 9.9 7.0 5.4 4.7 4.2 3.4 2.7
Kansas City, MO, ITC 15.1 12.5 8.8 6.8 6.0 5.3 4.3 3.4

38
LCOE benchmarks are highlighted in bold. As noted previously, we use the fixed-tilt systems for LCOE benchmarks from 2010-2015 and then switch to one-axis tracking
systems from 2016 to 2017
58
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.
New York, NY, ITC 16.7 13.8 9.7 7.6 6.6 5.9 4.7 3.7
Utility-scale (fixed-tilt)
Phoenix, AZ, no ITC 20.3 16.9 11.6 9.0 8.1 7.5 6.1 4.5
Kansas City, MO, no ITC 24.8 20.7 14.3 11.1 9.9 9.2 7.4 5.5
New York, NY, no ITC 26.5 22.1 15.3 11.8 10.6 9.8 7.9 5.9
Phoenix, AZ, ITC 12.9 10.8 7.6 6.0 5.3 4.9 4.0 3.0
Kansas City, MO, ITC 15.8 13.2 9.3 7.3 6.5 6.0 4.9 3.7
New York, NY, ITC 16.9 14.1 9.9 7.8 7.0 6.4 5.2 4.0

59
This report is available at no cost from the National Renewable Energy Laboratory at www.nrel.gov/publications.