E2E Networks Limited

CIN- L72900DL2009PLC341980
1st Floor, A-24/9, Mohan Cooperative Industrial Estate
Mathura Road, New Delhi-110044, Phone No. +91-11-4084-4964
Email : [email protected], website https://www.e2enetworks.com/

Date: May 30, 2024

Ref: NSE Symbol- E2E

Sub: Transcript of Analysts/Investor Earnings Conference Call Q4 & FY 2023-24

Dear Sir/Madam,

Pursuant to Regulation 30 of Securities and Exchange Board of India (Listing Obligations and Disclosure
Requirements) Regulations, 2015 and other applicable Regulations, please find enclosed herewith the
Transcript of Analysts/Investor Earnings Conference Call which was held on Tuesday, May 28, 2024 at
12:30 P.M. (IST) to discuss performance for the quarter and year ended March 31, 2024 (Q4 & FY 23-24).

The same shall also be made available on the website of the Company at https://www.e2enetworks.com.

This is for your information and records.

Thanking You

Yours faithfully,

For E2E Networks Limited

RONIT by RONIT
Date: 2024.05.30
11:54:45 +05'30'

Ronit Gaba
Company Secretary & Compliance Officer
Membership No. A59215
 E2E Networks Limited
Q4 FY24 Earnings Conference Call
May 28, 2024

Moderator: Ladies and gentlemen, good day, and welcome to the Q4 and FY24 Earnings Conference Call of
E2E Networks Limited hosted by the Valorem Advisors.

As a reminder all participant lines will be in the Listen-only mode and there will be an
opportunity for you to ask questions after the presentation concludes. Should you need
assistance during the conference call please signal an operator by pressing “*” then “0” on your
touchtone phone.

I now hand the conference over to Ms. Purvangi Jain from Valorem Advisors. Thank you and
over to you ma’am.

Purvangi Jain: Good afternoon, everyone and a warm welcome to you all. My name is Purvangi Jain from
Valorem Advisors.

We represent the investor relations of E2E Networks Limited. On behalf of the company, I
would like to thank you all for participating in the company’s Earnings Call for the Fourth
Quarter and Financial Year Ending 2024.

Before we begin a quick cautionary statement. Some of the statements made in today’s concall
may be forward-looking in nature. Such forward-looking statements are subject to risks and
uncertainties which could cause actual results to differ from those anticipated. Such
statements are based on management’s belief as well assumptions made by and information
currently available to the management. Audiences are cautioned not to place any undue
reliance on these forward-looking statements in making any investment decisions. The purpose
of today’s earnings conference call is purely to educate and bring awareness about the
Company’s fundamental business and Financial Quarter under review.

Now, I would like to introduce you to the management participating with us in today’s Earnings
Call and hand it over to them for their opening remarks. We have with us Mr. Tarun Dua –
Chairman and Managing Director and Mrs. Megha Raheja – Chief Financial Officer.

I now request Mr. Tarun Dua to start with his opening remarks. Thank you and over to you sir.

Tarun Dua: Thank you Purvangi and good afternoon everyone. Welcome to our earnings call for the fourth
quarter of the financial year ending 2024. Welcome to over earnings conference call for the

Page 1 of 16
last quarter of the Financial year ‘23-24. Before we begin with the performance for the fourth
quarter, and the Financial year ‘2024. Let me give you a very brief introduction about the
company for some of you here who might be new to the company.

E2E Networks was founded in 2009, and is amongst probably one of the largest public cloud
players in the cloud GPU market in India today. We have been providing cloud services since
2009 to various startups and businesses. And our primary focus is on workloads of artificial
intelligence, machine learning and for applications developers, enterprise customers, running
ERP and CRM solution, and we operate within the framework of infrastructure as a service
delivery model on cloud computing. Which primarily consist of like the workload management
for CPU, GPU, storage solution, containerization, virtual firewall, a lot of cloud native services
like SKUs services, load balancing services and more recently, like the underlying infrastructure
for artificial intelligence for training, and inference and deployment of model endpoints.

So we listed on NSE emerge in 2018. And more recently, we transitioned over to the main board
of NSE. Over the course of our evolution, we have significantly expanded our customer base,
and currently we service nearly 3000 active customers. We have our major facilities operating
out of Noida and Mumbai where we use the underlying third party data centers. A more
comprehensive list of product offerings and solutions of our cloud platform are accessible on
our self-service platform at myaccount.e2enetworks.com.

So, in the current world we have increasingly seen that artificial intelligence, machine learning,
and generative AI domain are the major new workloads, which are kind of like almost 4x to 5x
the typical size of CPU workload. And increasingly, like what we are seeing is that like, there is
going to be like a major shift of most of the workloads towards like centered around cloud GPU,
rather than the traditional CPU bound workload.

So, what we have seen is that, this presents a very substantial opportunity for us to kind of
start working at par with our competition, because frankly, everyone is trying to build at the
same time and figure out what would be the path forward for the enterprises and for the entire
set of organizations, how to adopt the cloud GPU. And multiple approaches, including that of
people working with hyperscaler, people working with cloud GPU platforms like us, which
provides a lot of underlying software infrastructure are being used. So we have seen as a
national imperative also, from the announcements that the government has made regarding
artificial intelligence and the support that the government is providing that, one of course that
the data center capacity required for artificial intelligence is going to kind of, exponentially
grow over next five years or so. And that data center capacity is going to be applied majorly by
AI/ML workload.

Now, we feel that having been in the field since 2019-20 onwards, we have an edge in terms
of our software, and listening capabilities for AI machine learning and generative AI workload
and we have historically been providing very high value compared to the pricing we charge.

Page 2 of 16
 And very, very centered on delivering value to our customers. So, today primarily what we see
is that, we are very, very focused once more on the startup ecosystem, where the new age
AI/ML and generative AI startup are increasingly working on solving problems using Cloud GPU.
And then basically, we also see a lot of demand from higher education and research. Apart
from the traditional workloads that we continue to help the SME companies with. And in the
near future obviously, our goal is to kind of expand this universe to include the workloads
coming in from larger enterprises as well. Now, we have been very competitive on our pricing
compared to the value we provide to our software and services. And this continues to remain
very, very attractive to our customer base.

Now, we feel that we will continue to be able to grow at a higher than industry average growth
rate over the next couple of years. And so it’s been a long while since we have transitioned
over to the self-service model, and the percentage of customers being served from our self-
service, public cloud platform where we provide a lot of software support, which is very, very
similar to what the hyper scalar cloud platforms do continues to serve us well. And in the
future, with a very generous dosage of the software being built, targeted at the data site base
and the AI/ML researchers and the gen AI company will continue to expand our offering and
our software to help people utilize the workloads on our infrastructure platform.

So during the last quarter of the Financial year ‘24, we have made very rapid progress on many
new features and services. And we will continue to release those new features on our AI/ML
platform called Tir within the workload environment for the AI/ML workloads, where we take
care of the ability to do all four major things. So the training, the inference, the model inference
deployment, and the ability to build data pipelines in a scalable fashion. And we continue to
add additional scalability features over there.

So, we have also kind of like integrated Tir better with over infrastructure services platform. So
you have the ability today to launch raw operating system environment using containers on
top of the paid platform that allows you to kind of have a very similar platform to what people
are used to on the my account itself. And it’s very, very seamlessly integrated with all the
software features for using AI/ML. So we are additionally working on compliance and security
features and monitoring features for our entire platform. And we continue to innovate on
storage technology for our infrastructure services platform. And broadly, we are very confident
of keeping up with the industry needs in terms of various verticals. And we are very happy to
note that we are being recognized as an AI first cloud GPU infrastructure platform in India with
technical capabilities, which are at par with the very best. So now, I would like to hand over the
call to our CFO, Megha to briefly talk about some of the financial and operational highlights of
the quarter under review. Over to you Megha.

Megha Raheja: Thank you Tarun and good afternoon everyone. Let me first start by giving you some of the
key financial highlights. I will summarize the performance for Q4. For Q4 FY24 the revenue from
operations stood at around INR 29 crore, which witnessed a growth of 67% on year-on-year

Page 3 of 16
 basis. EBITDA for the quarters is around 15 crore, which witnessed astounding growth of 68%
on year-on-year basis. And net profit was reported at 4 crore with a growth of 39% on year-on-
year basis.

Coming to financial year ended 2024, revenue from operations stood at INR 95 crore
representing a growth of around 43% year-on-year basis. EBITDA is 48 crore, representing a
growth of 45% on year-on-year basis, and a net profit of 22 crore, which witnessed a growth of
around 121% on year-on-year basis. That concludes the update for quarters and we can now
open the floor for question-and-answer session.

Moderator: Thank you very much. We will now begin the question-and-answer session. The first question
is from the line of Parikshit Kabra from Pkeday Advisors LLP. Please go ahead.

Parikshit Kabra: I just wanted to understand, how do you think this will play out as the competition in this space
increases, we know that Yota has already entered the space and has managed to create a
relationship with Nvidia as well where they are getting the chips. So my question has two parts,
one is how does your relationship with Nvidia exactly work and how do you make sure that you
continue to get the access to the chips and number two is, how do you think you will be able
to manage to keep your right to win as the number of players in the space increase?

Tarun Dua: Okay. So, first part about like whether people will get access to the chip, like what we saw in
last 18 months have the kind of shortage of chips, those are like the kind of events that
happened like once in a while. Now, there was obviously like 10x growth in the overall cloud
GPU consumption and the GPU consumption globally. So that obviously resulted in like, short
term shortages. Now, once that shortage is there, once the demand is clarified then obviously,
the foundries and the entire ecosystem that is at the back end of the major chip makers geared
towards the new demands, and that takes a couple of quarters, which has already happened.
So we don’t feel that in the future, that there is going to be that sort of shortage. And you will
need like very, very special relationships to kind of like move ahead of the queue. Now, that
being said, like we continue to enjoy a very good working relationship with all of our backend
vendors, including and Nvidia. And the way we work is that like, basically we bring our back
end vendors into the conversation with our customers whom we are talking to, and helping
them with their workflows. And it is to kind of like help validate any solutions that we are
building for our customers. Now, that relationship continues as well. And that relationship will
continue in the future as well. Now, regarding like heavy competition, like I wouldn’t like to
comment on any individual competitors. But for the past 15 years, as a business, as a company
we have operated in what is called a Red Ocean. So, we have never shied away from
competition, we continue to build up our portfolio of software, and services and solutions with
a view to develop a better understanding of the workloads that our customers are bringing to
us. And that insight into the workload, that ability to understand the need for the workloads
and implementing those in the software. That product differentiation is what creates long-term
value for our customers. And we’ll continue to do that. So as you continue to get the adoption

Page 4 of 16
 on your software and you keep getting the feedback from the customer, you keep building
based on that. And that creates the long term value and differentiation for any business in
software and cloud industry. And we will continue to do that.

Parikshit Kabra: So, just a follow up on that, first of all thank you for a detailed answer Tarun. But as a follow
up on the part with the chip shortages so your opinion is that, in the next couple of quarters,
the queue for getting these GPUs will no longer be a concern, the supply constraint will be gone
in a couple of quarters, because I got a sense that is going to be.

Tarun Dua: There is always going to be some queue. But it’s not going to be the longest queue that we
have seen in the past of that type okay, like we are looking at 52 weeks, 56 weeks, 60 weeks
and so on. That is not going to be the future. A normal queue and mostly because end of like
overall, that’s the nature of integration required to deliver the chip in a form that can be utilized
by cloud operators. So nobody buys like raw GPUs anymore like we are always packaged in the
form of like HGX conserve or other server formats that are coming up. The entire integration
takes time. And like people have to understand the entire solution before they can implement
it. So that is the part that takes time so, I don’t think like it is. So, all the moving parts have to
be in sync for the solution to be built for you and then deployed at your location. So all of that
part takes time but like I don’t think like there is going to be a long waiting to queue in the
future as well.

Parikshit Kabra: Got it, so we don’t have a format priority relationship with Nvidia and other vendors. It’s a
informal network as such, is that a fair statement then?

Tarun Dua: Let’s see. Like if you go and search on the website of the major vendors like we are obviously
listed as a cloud service provider partner and from the perspective of like, all the support that
we have demanded and got, I feel that like it’s a type of a relationship without putting any
additional formal requirements to it, it continues to be a great relationship.

Moderator: Thank you. The next question is from the line of Ayush Bansal from Emkay Global. Please go
ahead.

Ayush Bansal: So, looking at numbers, there is a 185 crore CAPEX in FY24 which was 35 crore last year. So,
what is the revenue potential for this CAPEX and is this reflected in the MRM as of now?

Tarun Dua: So some part of the MRR reflects this CAPEX obviously and what we believe is that, of course
like the CAPEX kind of like reflects into MRR and number of different SKUs and each SKU
obviously based on the amount of value added software, the nature of the length of the
contract and so on. So, depending on many factors like that, MRR realization is differentiated,
but regardless of that, based on our current understanding, we believe that like the MRR
potential with the current CAPEX that we have already done would be of the order of anywhere

Page 5 of 16
 between like 14 to 16 crores from the current MRR number like we should be able to achieve
over like next couple of months.

Ayush Bansal: Okay, got it. And is there any further plan for CAPEX during FY25?

Tarun Dua: It’s a continuous process like as we continue to project the demand coming in from our
customers, we continuously make the expense on CAPEX, so there is no CAPEX done and
dusted ever in our cloud industry. So, we are always buying.

Ayush Bansal: Okay. And can you help us quantify that number, if you have anything in mind?

Tarun Dua: In an ideal world, like we would like to end another $100 billion or about Rs.800 crores over
the next 12 months.

Ayush Bansal: Okay. And how do you see that to be funded?

Tarun Dua: It’s will have to be a combination of things including like financing from banks and leasing
partners and vendor financing and some equity raise.

Moderator: Thank you. The next question is from the line of Abhishek from ABC Capital. Please go ahead.

Abhishek: In the November concall you had guided for a growth of 40% to 50% top line. So, can we get
that next year, the revenue will be around 140 crores and you have given that margin will be
around 30%. So, we can guess that the bottom line will be around 40 crores, is it correct?

Tarun Dua: See, I leave that task typically to an analyst, we do our best. So, like obviously we kind of have
some sense of projected demand and we have some sense of like how much capacity we need
to put up, now that conversion happening over like predictability of that conversion. So, if we
can not promise that, but obviously we try our best to kind of like convert the capacity we have
to revenue as soon as possible.

Abhishek: Fine. One more single question, like can we consider that since you are providing background
support for all the software’s which these startups are doing. So, can we consider E2E Networks
like a proxy for all the new age computer based startups new age startups that are coming, can
we compare like that, is it like that as the startups go, we can think that our regulation company
also will work and…..

Tarun Dua: I can’t clearly the question. So, is your question is like overall, if the AI/ML GenAI business
grows, then are being to see a lot of growth because, that was like a lot of GenAI startup, the
short answer to that is yes obviously, there are no answers to, basically like what those startups
are doing and what are their workload requirements and how many of them are coming to us,
and so on. But like broadly from a perspective of where the market seems to be going, in
general if you are seeing a lot more AI/ML adoption, overall, both in the industry as a little bit

Page 6 of 16
 startup as well as into higher education and research, then yes, that growth is in a sense like a
proxy for our growth as well.

Moderator: Thank you. The next question is from the line of Pankaj Shah from Dinar Wealth. Please go
ahead.

Pankaj Shah: So, wanted to understand, since your demand is so high for the AI/ML workloads, how quickly
can we deploy this CAPEX, this 150 crore CAPEX which we have done?

Tarun Dua: For the CAPEX done, is already deployed on our cloud platform, how quickly can we sell is like
something that we do not make a statement on, that is something we like to look in the past
once we have done it. So, that’s the way we wish to operate.

Pankaj Shah: So in past, you used to guide that over a quarter or two quarters, we used to sell the capacity.
But the question is like, if the demand is so high, do we see that change?

Tarun Dua: That is always the guidance we give, there is obviously like a cycle to kind of like selling to the
cloud customers. So, where obviously you continue to stay in touch with the entire ecosystem,
continue to work with people on POCs, continue to kind of like guide them on how best to
optimize their workloads on our cloud platform. So that is the typical explanation of the cycle
of a quarter or two. But obviously, like having a shorter period, or having a longer period is
always possible. So you can’t predict the future for like let’s long term future you can predict
and yes, over the next five to 10 years, we are going to be explosive growth in the AI/ML field.
But can you predict like a quarter-on-quarter or say that like next two quarters, whatever is
going to happen, that is something which is very, very hard to predict, at any given point of
time.

Pankaj Shah: Right. So the question was not from a quarter-on-quarter perspective, it was just that if the
demand is high, and are we seeing the changes in the deployment?

Tarun Dua: The demand is only going to increase as the industry shifts to more cloud GPU centric
workloads. That demand is only going to increase in the industry. And obviously, it is not
possible to predict like, what’s going to happen in the next one quarter or next one years, but
it is possible to kind of make a view over a medium term or long term of five to 10 years and
say that yes, this is a broad industry trend with the entire industry thing to agree on.

Pankaj Shah: Okay, got it. And sir how should one look at the asset turn, so earlier we used to get a 0.7x over
a year. So does this increase with time or how should we look at?

Tarun Dua: See, it’s very, very difficult to kind of, these are all averages, in the sense that like the asset turn
also relies on a number of things, the period for which the customers are using the services,
so, then there is more scarcity in the market, people want like longer contract period, when

Page 7 of 16
 there are less scarcity, then obviously people want like shorter periods, but that doesn’t mean
that they are going away tomorrow, day after or next month. But they also tend to end up
paying a certain percentage more for shorter periods. Now, that being said, like there is also
the nature of SKUs, so for the same underlying hardware, assets, like there are a number of
ways in which the assets get monetized due to the nature of the software capabilities as well.
So over there, for each different SKU, the asset terms could vary from all the way from point
0.5 to even like 2. So what is the final blend we are able to achieve like, is again something we
can look back into the past and say that, okay this was what was achieved but, there is a lower
bound for that which we can predict, which is like somewhere, anywhere between say 0.55 to
0.6. But that’s the lower bound. But, on the other hand, like there is no upper bound in a
software driven cloud, like where, what sort of capability you can bring to the table to achieve
like an asset turn. So normally, if you ask me, that was not how we look at our business, we
look at our business as a technology business, very focused on figuring out the various industry
solutions. And over there eventually like, the asset turn doesn’t matter, because it’s like, very,
very capability and technology and software driven business and it’s not about being the
cheapest on the market. It’s about creating enough value for the customers.

Pankaj Shah: Okay, got it. If I got it it’s correct, you said 800 crore CAPEX plan for FY25 right?

Tarun Dua: Ideally, in an ideal world, like I believe that might be something we might want to do.

Pankaj Shah: Okay. And a few on the bookkeeping side, the line item which is the purchase of consumables
and services, what exactly is this item?

Tarun Dua: See primarily like the major expenses that we have are on the data center side as you are aware
that we use third party data centers, the major amount of services purchased are related to
data centers services that we purchase.

Pankaj Shah: So is that a power or anything else?

Tarun Dua: It’s all like a combination of power, base, bandwidth and certain other smaller items. Just as a
part of this. Software purchase is also there, so some of it intended for the customers where
there is a direct one-to-one relationship between services purchased and services delivered to
the customers, the examples of that would be the licenses, of the licenses that we delivered to
our customers. And then there are obviously like certain other services that we also procured
in the normal course of planning….

Pankaj Shah: Right. So what I was trying to understand is, majorly we should look at expense as a fixed or
more of a variable?

Tarun Dua: So the services would be like in the same proportion as the revenue or a way that they will
increase with the revenue in a certain incremental percentage fashion. Now again, it’s the

Page 8 of 16
 question of like, there is a broad number that is there and broadly that percentage will kind of
like as the volume increases, like go down slowly.

Pankaj Shah: Okay. Sir what exactly goes into the other expense, because I am sure the lease rentals are
being calculated below the EBITDA with IndAS 116. It majorly flows through depreciation and
interest. So what exactly it does?

Tarun Dua: Let me ask you to repeat the question that get Megha to answer some very specific finance
questions you are asking me because, I come from engineering background now completely
into the business part. But, I let Megha answer your question if you can repeat your question.

Pankaj Shah: Sure. So, one thing on the balance sheet side, the right to use asset which we have created, I
assume that that’s mainly on the data center lease visibility which we have and if I come on the
P&L side the lease expense will be covered mainly through depreciation and interest. So what
exactly are other expenses if I’m understanding this correctly?

Megha Raheja: Sure. So in other expenses we have payment gateway charges, then membership subscription,
legal professional fees, these type of expenses and advertising and promotion, these expenses.

Pankaj Shah: Okay. And this right to use is mainly for the least visibility which we have, data center?

Megha Raheja: Lease is separately capitalized as per IndAS.

Tarun Dua: So, the data centers were not lease, data centers are rented in a not strictly in the question of
trend, but data center services like they have their own classification. But on the other hand,
like leasing relate to the equipment and this is all lease to own kind of equipment.

Moderator: Thank you. The next question is from the line of Ratan Mehta an Individual Investor. Please go
ahead.

Ratan Mehta: Hi, Tarun I have been your customer for many years of your company. I came to know of you
guys about six, seven years ago and since then, you have given terrific service. And a few years
ago, I had the opportunity to invest. So, I did, so I have a couple of questions. If you don’t mind,
I’ll just dumb it down to myself. Down the business model, which is the civic buyer box, and
this box is your standard CPU RAM, you take pieces of this box, and you rent it out and get
more rent than the cost of this box. Then there’s this new box, which is a GPU, and you rent
this box out again, with the same model, you are going to get more money for it. And then
there is a third box which is cloud storage, which is just hard disk, and you get money for that.
So this accurately kind of describes the economics of the business?

Tarun Dua: Okay. So in the sense that, where you are saying that you have combined software with the
boxes. And ultimately, you are selling a box is that the definition of the business, I somewhat
disagree to that, because ultimately you are running a workload. But the way to run the

Page 9 of 16
 workload, how to make it more efficient that is what the people pay for on the cloud.
Otherwise, everyone would be kind of like running like on prem, and kind of like company-
locating on their own and managing on their own. So, as the world becomes more complex,
like a typical cloud operator is doing something like a couple of 100 activities on the side of
security, on the side of reliability, on the side of scalability, on the side of like being able to, of
course the part about being able to utilize the resources on the tap, that’s the whole other
infrastructure. So, typical number of activities that a cloud operator is doing for you, which you
cannot very easily replicate in a business is numbers around a couple of 100. Most of these
being done in software’s also, a lot of these activities are done as standardized processes that
take root over decades of experience from the personnel and codified both into software as
well as into the processes for the people. And also for basically, what you are demanding from
your vendors, and so on. So, very, very when you say that dumbing it down like really dumping
it down, that ultimately you are selling a box.

Ratan Mehta: No, I especially get the nuance in that.

Tarun Dua: Box is what we are selling, so potentially you could kind of, so it’s a whole ecosystem that has
to be built, so it’s not something easy to replicate and say that okay, let me deploy some
software onto a box. And I can be an equivalent of like E2E Networks or one of the competitors
who has been around for a couple of years to kind of achieve that.

Ratan Mehta: Thank you so much for really explaining that appreciated. I didn’t mean to simplify over the
middle. But I really appreciate your clarification. The next question based on this little analogy
is, when you are using debt to grow, you are now putting a lot of the debt towards GPUs in the
hopes that AI and GPU game is going to be where you make the most amount of money. And
the CPU game, you may have reduced, I’m not sure about this, but you may have reduced the
amount of debt you take to grow in the CPU route, so is that a risk to the business and how
would that layout?

Tarun Dua: Not much really so, the CPU side also continues to be helped by both the tailwind. So one is
that like anyone who is buying that GPU is not necessarily just using the GPU, they are also
necessarily buying the other parts of the cloud ecosystem so which also helps over there. And
second, the organic growth of the existing set of customers who are very focused on the GPU
workloads, we continue to pay attention and continue to pay great focus to that and continue
to build software features for that. And what our very sincere belief is that like, when people
are looking at a cloud, they are looking at it holistically, they are not coming in just for the GPU
capabilities alone. Now, during a scarcity era, that could have been the case where people are
like, okay, all I need is the GPU cloud, I don’t need anything else. But typically, for any
developer, data scientists, AI/ML scientists they require the entire cloud ecosystem, and not
just the GPU side of things. Now, that being said that like, yes the CPU acquisitions do get vast.
So for an example let’s say, someone like an E2E Networks spends let say 200 crores,

Page 10 of 16
 supposedly, all of it into GPU. But I’m assuming that like, it could also have a 10% portion, which
is actually going towards like storage and CPU side of the cloud as well.

Ratan Mehta: I understand, my concern only comes from because obviously you are in the cloud space. The
noise around you, as the CEO would be everyone around you saying GPU, GPU, GPU. And the
risk in terms of CPU versus GPU is coming down to watch how you, what noise you heard. And
that might bias you a little way which is where I’m asking the question from. So my concern is
that, five years from now, three years from now, or two years from now, it’s discovered that
there aren’t that many applications in AI, which is a possibility. A lot of this debt then cannot
be repaid with revenue, because perhaps the GPUs don’t pay off, that debt. Do you see that
future playing out at all or how you mitigated that risk?

Tarun Dua: No, so one is that like, of course we are not a dumb CPU or dumb GPU rental company. So in
the sense that like, we continue to be a software driven business, and what the investment
represents is the capacity to service customers. In an ideal world, as we would like to have
unlimited deadlines to kind of expand the capacity in an unlimited fashion, and not even have
to worry about where the money is ideally coming for on the hardware side at all. But that
being said, I don’t see there is any risk, the process of going to the GPU is irrevocable. The kind
of exponential computing offered by GPU is kind of like really changes the era in which we are
operating. So, I don’t think we are going back to the era of people kind of investing more than
25%, 30% on the CPU side, but major investment in the entire IT industry are going to be on
the heavier GPU compute side. And there would be like, definitely more than one vendor on
the GPU side as well. And there would be multiple aspects, so if you look at all the CPU
companies, they are also getting into the GPU game. That tells us that yes, CPU is going to
continue to be there for a very long time. But the majority of industry focus is going to be
solutions which are built on top of the GPU.

Ratan Mehta: Thank you. The next question is from the line of Kshitij Saraf from Tusk Investments. Please go
ahead.

Kshitij Saraf: I have a couple of questions. Firstly, how fluid is the CAPEX plan for you, so are you looking to
spend it mostly on Nvidia GPUs, or are you also following developments around Microsoft
coming up with its own set of copilot PCs and sort of PRM also comes into the picture over
there. So, just some sense of how you are looking at this whole vendor landscape and adjusting
your CAPEX or thinking about it.

Tarun Dua: Sure. So, in the short term all our GPU investments that I see are definitely all going to Nvidia.
Now, there is obviously no exclusivity as such in the sense that tomorrow if somebody is able
to bring more capable check for any specialized used cases, we would definitely evaluate that
independently and take a call on that at that point of time. So, like we are very, very open to
kind of be very flexible about how do we do over CAPEX based on the customer demand. So,

Page 11 of 16
 it’s obviously not all going to be made in one shot, it would be spread over some word. And
kind of that’s the way we have always operated.

Kshitij Saraf: Understood helpful. Secondly, on the fundraise, on the equity bit, what are the kind of
timelines we are looking at and are we looking to deploy the entire bit at once, once the
approval comes through?

Tarun Dua: So yes, the ability to kind of like consume capital for us is not in doubt. Now, we are working
on it, like over like next couple of quarters, I guess.

Moderator: Thank you. The next question is from the line of Ameya Pimpalgaonkar an Individual Investor.
Please go ahead.

Ameya Pimpalgaonkar: I just have a couple of questions. And some of my questions have already been answered. So,
I will not repeat those. I just want to take your attention to the recent outage that we
experienced at one of your data centers. So, I just wanted to understand from you. Has it
caused any kind of impact, let’s say in terms of, of course maybe not in financial terms, but in
terms of customer experience of sorts, maybe it could be nice if you start also publishing some
of your customer experience course, maybe not this quarter maybe coming quarter. So this is
my first question. I will come to the second question after this.

Tarun Dua: Sure. See related to the outage like we have kind of completely and transparently shared the
entire timeline. And what exactly happened with all of our customers. So obviously, all the
details like are basically the primary responsibility we have is towards the customers. Now what
happens in case of a outage at large where there is a complete blackout, those kinds of
blackouts, like even when we were talking to our vendor, they said that seen like, this kind of
a blackout like once in a decade, kind of a thing. It’s not that this happens every day, of course,
there is always a probability of something like this happening. And it does happen to all the
major providers at some point or the other, impacting, one or more of their facilities. So, what
happened in the outage was that like, obviously there were a complete blackout. Now, there
was at the level of our third party vendors, there was a miss from their side, where they kind
of continued try to power on without waiting for the UPS to be charged fully. Now, this created
issues for our recovery, because some of our equipment failed because of the surging power
and going away of people like a couple of times, and we had to kind of throw away some of the
equipment damaged due to that. And then on the software side, you have to take care of
verifying the impacts on a per customer basis, one-by-one. Now, typically, if you look at any
operations team, the operations teams are decided, are kind of there to have the ability to deal
with some level of surge, so let’s say you have a couple of 1000 customers, typically in a week
maybe 5% or 10% of those customers come back asking for personalized support and which is
definitely done. Now, in this type of an outage where the impact runs into a very large
percentage of the customer base. Obviously, your normal support processes do not work,
everyone on calls cannot be answered although text messages and emails can be answered as

Page 12 of 16
 per the queue. And based on, some of the priorities where the customers come in and say that
look this is very critical to my business. Can you prioritize it, obviously we do the –50:05 kind
of go and solve those customers problems first. And obviously, the goal is to kind of, service all
of our, our entire customer base uniformly over there. So, yes this was already explained to all
the customers, this was a black swan event that happened. And most of our customers were
very understanding about it. Some customers had, even some other issues, which they opened
up about that, this was a black swan event, for sure. But I have other problems. So that also
allowed us to engage with many of those customers and solve many of the other problems as
well. And so we have, two major facilities now in Noida and Greater Noida, and this happened
at older facility. And our plan over next couple of years is to kind of slowly migrate without
disruption all the customers in this facility, and obviously prioritize those customers who are
kind of willing to put in the effort today to help them migrate over to our new facility. And we
are also going to be putting up like a 200 gig link between the two facilities to facilitate that.
The new facility is like very, very new. The plan was always to put up a very long broad link
between the two facilities to link them together. But yes, like we have prioritized this and kind
of we prioritize solving the problem for all the customers who could be brought up. Now, if
that power outage had only happened on the power outage level, and kind of, so it’s a lot of
ifs and buts and what could have been done, should have been done. That’s all in the past now,
but, I would like to point out that, in case that the third party vendor had trained their processes
better, we could have avoided a world of pain for us and for our customers, by putting up the
power only after the UPS was charged, instead of trying to bring it up couple of times, which
damaged some of our equipment. And of course, we had enough equipment on site to be able
to kind of continue the operation post replacement for all the majority of our customers. Yes,
we spend a lot of time in verifying, or each of our customers whether their services are up and
over couple of like 24, 48 hours that we continue to call our customers and make sure that
everyone’s services are completely up and running.

Ameya Pimpalgaonkar: Understood, thank you that was helpful. And my second question Tarun is on.

Tarun Dua: Let Ameya continue with the second question, technically he has asked only one question.

Ameya Pimpalgaonkar: Thank you so much, I appreciate. I had to ask you this, because you are doing a lot of value
addition in terms of your software ecosystem as well as at a network level for example, you
have TIR framework, you have InfiniBand, that you have recently deployed, both 1.6 Tb and 3.2
Tb. Just wanted to, pick your brains, what nature of advantages let’s say this InfiniBand is
bringing you compared to the peers, when we talk about embedding a value, when we sell a
contract to a customer, what sort of value are you able to bundle in your offerings because
these small things are, not coming to the front and I just wanted to understand how you are
looking at this, when you go to a customer, what kind of value are you able to attach to these
things for now that you have. Thank you.

Page 13 of 16
Tarun Dua: Sure, thank you Ameya. That’s a important question to address, in cloud what we are doing is,
like I said that we are doing hundreds of activities and integrating like hundreds of different
services based on open source and hardware and the processes our team to be able to train
that. Now, it is to what end so one end is that like, we are always providing the latest greatest
technology to our customers at all times. So that is always the goal, now the second part like
you asked specifically about 3.2 Tbps kind of InfiniBand so, the overall idea is that, what is
difference between the old world of the CPU and the storage and the new world of CPU plus
GPU plus very high speed panel file system kind of storage and so on. Is that like earlier what
you were doing was that like, the nature of your software was dependent on the abstraction
that you provided in your data, which was like you were choosing to store the data, which was
by its very nature, like key value payers or relational or some sort of structuring of data was
required. So your data was going into structure databases, and the software was limited by
what you were storing in terms of the structured data. Now, structured data in any organization
is actually only about actually 5% or maybe even less of the amount of data that any
organization produces. All the other 90%, 95% of the data is actually unstructured data. Now,
if you look at the new world of AI, where GPU has almost 100 times the processing capability,
now that 100 times that processing capability is actually going to be used for the remaining
95% of the data that does not get captured in structured database format. Now, which means
that like earlier you were designing your storage and network and your entire internal
bandwidth infrastructure for being able to process like very small amounts of data. Now you
actually are building for like being able to handle like 20x, 100x or maybe even 1000 times the
amount of the data in motion to be able to process it on like very fast, high speed, GPU
processes. So that’s why we are building all the technology around like why do you need a
bigger network. Why do you need a 400 gig or 800 gig Ethernet and why do you need a 3.2
Tbps InfiniBand is to address the data in motion at these scales. Because these are the scales
required for dealing with organizational data which is unstructured, so which is primarily used
to be classified into say voice or natural language, or LLMs as they are called today in GenAI
term or video or images. And tomorrow, that nature of data is going to become multimodal.
And like, we may not need to store all of that data, only the weights and biases kind of like
generated through that data and to vector databases and so on. But yes, like you still need to
be able to handle like a lot of data in motion. For that, you need to kind of like completely
upgrade the way in which you think about technology.

Moderator: Thank you. The next question is from the line of Shashank Rastogi an Individual Investor. Please
go ahead.

Shashank Rastogi: I have just one query. So my question is that in the last quarter, your deposition part increased
a lot, is there any change in your accounting policy. If it’s the same currently.

Tarun Dua: Let Megha answer this question, and after this question from your side, we’ll take like one
more participant question. And then we would end the call. But let’s have Megha’s answer to
your question and then we will take one more question and then we will end the call.

Page 14 of 16
Megha Raheja: So there is no change in accounting policy as regards depreciation. The depreciation was
changed from April ‘23, where we have changed from WDV to straight line method in line with
the Companies Act. That answers your question?

Shashank Rastogi: Yes.

Moderator: Thank you. The next question is from the line of Parikshit Kabra from Pkeday Advisors LLP.
Please go ahead.

Parikshit Kabra: I have two questions. First is to understand the kind of customer profiles that we have. I know
you mentioned that startups are our focus area, but are we also seeing that as the operations
of our customers scale, they migrate to quote unquote, the bigger players like AWS and Azure
for stability or security reasons or are we not seeing that pattern as of now?

Tarun Dua: Okay, so the nature of our customers like primarily is the four different buckets today. So one
is of course, like we have the SME segment, which continues to be our main the bulk of our
customers, then we have the higher education and research category, then we are back with a
bang with working with a lot of startups today. And fortunately, we also work with like an auto
flex partner ecosystem, both in India and outside India. And they also bring in like a lot of
customers. And to answer your question about like do customers migrate. I don’t think there
is any imperative to migrate today as such among the customers from anywhere to anywhere.
The nature of workloads has become multi cloud, so it feels that like in a single organization,
multiple cloud is given. So nobody is trying to consolidate into a single cloud platform anymore.
So there is no imperative to migrate the workloads, so wherever the workload are running
most efficiently by any team who is running the workload. They usually continue to run over
there.

Parikshit Kabra: Got it. My second question, if I can get that in.

Tarun Dua: Please go ahead Mr. Parikshit.

Parikshit Kabra: Thank you. So, I was just trying to understand what is the split of our revenue from following
foreign customers right now and how are we, what is the GTM for that?

Tarun Dua: I don’t think we are doing a formal split right now. But currently, like our focus continues to be
in India, although we would obviously not refuse any customers coming from outside as well.
So from a focus outside your perspective, like majorly we are focused on the EMEA market and
the US market. And hopefully, we will start focusing on the Southeast Asian market as well, but
the bulk of our revenues continue to come from India.

Moderator: Thank you. The last question is from the line of Abhishek from ABC Capital. Please go ahead.

Page 15 of 16
Abhishek: Just one question, what is the team size right now and is there any plan to increase net count
next year and if yes, then by how much?

Tarun Dua: Broadly, we have claim to somewhere between 140 to 150 people in the team. So
approximately about 150 and we continue to grow the team organically. So, specifically
wherever we have access to talent, we continue to do that. And it’s going to be an organic
growth, so we don’t have any kind of a super formalized plan, we have seen a gap. And based
on those gaps, we continue to fill those gaps both in our technology team and our marketing
team, as well as the operations team.

Abhishek: So, no significant increase in the employee cost next year?

Tarun Dua: So again, it’s a question of like, so we continue to operate in like startup mode, we are
obviously, wherever we find talent, so we are happy to kind of entertain the talent and figure
out like how they can help us and the doors for hiring are never closed at E2E ever.

Moderator: Thank you. As there are no further questions, I will now like to hand the conference over to the
management for closing comments. Over to you sir.

Tarun Dua: Thank you everyone for participating in this earnings concall. I hope we have been able to
answer your question satisfactorily. If you have any further questions or you would like to know
more about the company, please do reach out to our IR managers at Valorem Advisors. And
thank you very much.

Moderator: Thank you. On behalf of E2E Networks Limited, that concludes this conference. Thank you for
joining us and you may now disconnect your lines. Thank you.

Page 16 of 16