Speaker #1: Participant lines will be in the listen-only mode, and there'll be an opportunity for you to ask questions after the presentation concludes. Should you need assistance during the conference call, please signal and operate by pressing * and then 0 on your touchstone phone.
Speaker #1: I now have the conference over to Mr. Vanessa Fernandez from Investor Relations team at AdFactors PR. Thank you, and over to you, Vanessa.
Speaker #2: Thank you, Myron. Good evening, everyone. On behalf of E2E Networks Limited, I would like to welcome you all to the earnings conference call for the second quarter of FY 2027.
Speaker #2: Today on the call, we have the task from the management, Mr. Tarunduwa, managing director of E2E Network, and Mr. Nitin Jen, chief financial officer of E2E Network.
Speaker #2: We will begin the call with brief opening remarks from the management, followed by a Q&A session. All participant lines will be in listen-only mode during the management's remarks, and the flow will be open for questions thereafter.
Speaker #2: Please note that certain statements made during this call may be forward-looking in nature. Such forward-looking statements are subject to risks and uncertainties that could cause actual results or projections to defer, materially from those statements.
Speaker #2: E2E Networks will hold no responsibility for any actions taken based on such statements, and undertakes no obligation to publicly update these forward-looking statements. I now hand over the call to Mr. Tarundua for his opening remarks.
Speaker #2: Over to you, Tarun.
Speaker #3: hi everyone. Thanks for joining us today. In our second quarter earnings call for FY 2027, this has been another strong quarter for E2E Networks.
Speaker #3: Where we have continued executing similar to the previous quarter, and this is still, like, day zero of AI for us, as far as we can see.
Speaker #3: And I would like to kind of, like, give a brief update on, in the previous quarter, so like, we signed, like, a nearly 1,000 crore deal with, like, a sovereign AI company, for the Blackwell cluster.
Speaker #3: Which lasts up to, like, June 2029. So this is something that improves the long-term revenue visibility and validates E2E's ability to transition the deployed capacity from a ASU model to a longer-term commitment.
Speaker #3: So apart from this, like, basically, like, this quarter we have really focused on increasing the talent density within E2E, in all areas of execution.
Speaker #3: So we have, like, a new chief revenue officer, we have also added, like, a team to the founder's office, we also have, like, a field CTO to kind of, like, help our customers understand our solutions better.
Speaker #3: And all of this brings, like, a lot of net new experience from the industry in, like, various facets of the industry. So essentially, like, that strengthens the management and leadership teams.
Speaker #3: Where, like, it primes us for the next phase of growth for ourselves. Now, apart from the inside talent within E2E, like, we have also strengthened our strategic advisory team.
Speaker #3: So which now, where we would like to welcome Mr. Anandmoy Roy, as a strategic advisor. So Roy has been, like, a part of the tech ecosystem of India for a very, very long time, and essentially, like, he brings, like, a lot of deep-rooted knowledge of how AI is being applied into the new ecosystem that is getting built.
Speaker #3: And his experience and perspectives, like, would be really valuable in helping us identify very many opportunities. And this would help shape our long-term growth strategy as well.
Speaker #3: Now, we continue to invest very rapidly into our sovereign AI platform, both on the side of keys and on the side of Jarvis. And, like, what I believe is that, like, the platform is the key to our future growth.
Speaker #3: So where, essentially, like, that allows us to operate as a full-stack player where we kind of, like, manage all parts of the sovereign AI multiple layers for our customers.
Speaker #3: And that allows them to use our GPU platform with confidence for not just training but also inference and the upcoming agentic workloads. And this, of course, requires, like, some part of our GPUs to be used by our own teams rather than, like, being completely pulled out to end customers.
Speaker #3: And that, this is something that allows us to kind of, like, go forward in the future with boldness. So now, finally, about, like, the future growth, like, we have always maintained that, like, we will continue to be both aggressive as well as prudential in terms of our capex deployment.
Speaker #3: Now, we do not, like, give a speculative capex guidance, and what we prefer to kind of, like, inform everyone of is, when we have, like, sufficiently certified and solidified our deployment plans, so that's when we talk about, like, yeah, that's, like, this is the potential capex deployment upgrade.
Speaker #3: That we are working on. So like, once it certifies, that's when we talk about it. Now, I would like to hand over the conversation to our CFO, Mr. Nithin, he can talk about, like, the, some of the recent financial-related updates for the last quarter.
Speaker #3: Over to you, Nithin.
Speaker #4: Thank you, Tarun. Good evening, everyone, and thank you for joining us today. I would walk you through our financial performance for the second quarter, and first half of FY 2027.
Speaker #4: Q2 FY 2027 has been another strong quarter for E2E Networks. Revenue for the quarter stands as 2 to 38 million, which is up by 411% year-on-year and 43% quarter-on-quarter.
Speaker #4: EBITDA stood at 1,715 million, up 45% quarter-on-quarter, with margin expanding to 76.6%. EBITDA margin expanded by 141 basis points compared to Q1 FY 27.
Speaker #4: Profit before tax clocked at 971 million, compared to 586 million in Q1, an increase of 66%. Tax stood at 727 million, compared to 439 million in Q1, with tax margin improving to 32.5% from 28%.
Speaker #4: Diluted EPS stands at 3.48, compared to 2.1 in Q1. The performance is driven by contribution from the Blackwell GPU, better utilization, and higher average utilization across the GPUs.
Speaker #4: H1 FY 27 revenue is 3805 million, which is up by 376% year-on-year. EBITDA stands as 2894 million at 76%, which is 2.3 times for the full year FY 26 EBITDA.
Speaker #4: PBT stands at 1,558 million, compared to a loss of 2222 million in H1 FY 26. And tax stands at 1166 million, compared to loss of 163 million in H1 FY 26.
Speaker #4: Now, I would hand it over back to the moderator for the question-answer session.
Speaker #3: Thank you. We will now begin the question-answer session. Anyone who wishes to ask a question may place star and 1 on the touchstone phone.
Speaker #3: If you wish to remove yourself from the question queue, you may press star and 2. Ladies and gentlemen, we will wait for a moment while the question queue assembles.
Speaker #3: Participants are requested to kindly restrict their questions to one main question and a follow-up to maintain equal opportunity. We have the first question on the line of Keshav.
Speaker #3: The question to please go ahead with your question.
Speaker #2: The congratulations on the great quarter. Like, like, few questions. Like, when are we expected to live basically go live with our next B200s that were planned originally, and when we can expect the revenue to be started from that?
Speaker #4: Yeah, so the next lot of B200, which we have already announced, like, is expected to go live this quarter. So hopefully soon. And like, when it goes live, obviously we'll keep everyone informed.
Speaker #3: Oh, can you hear us, Keshav? Yeah, I see no response. We'll move on to the next question. We have the next question from the line of Bharat Gulati from Dalal and Brocher.
Speaker #3: Please go ahead.
Speaker #4: Yeah, hi, hi.
Speaker #2: Yeah, hi. Hi, hi, Tarun. So thank you for the opportunity. Just a question on the 1,000 crore contract in the current 75 crore MRR, how much would be the contribution of that 1,000 crore contract?
Speaker #2: And you've mentioned that the entire, that the B200s have been multi-year contracted for the order. So is this the existing B200s, or is it going to be both the batches of B200s?
Speaker #4: This is the existing B200s, which got converted from pay-as-you-go to, like, a multi-year contract. Now, exact contribution would vary based on the usage, but, like, obviously it's easy to calculate, like, close enough by looking at the 1,000 crore number and the time period.
Speaker #2: So has the entire, I mean, is the MRR completely based in from the 1,000 crore contract right now, or is there some incremental MRR still to come?
Speaker #4: See, like, I don't know what, how to answer that question. Like, so basically, like, you are asking question about, like, a future that I cannot predict.
Speaker #2: Got that, got that. Okay. And just on the gross margin, sir, we've been calling out that 85% gross margins is where we'll be, but we've significantly beaten that this quarter.
Speaker #2: So how sustainable is this 88% margin? And just on the operating costs also, I understand we've made some investment in people, but where do we see absolute operating costs moving going forward?
Speaker #2: You know, it's 18 crores for this quarter. So if you can just throw some light on that.
Speaker #4: I'm not sure about, like, the 85% number. Like, like, see, overall, like, we are looking at this as a 70% to, like, 70% plus kind of, like, an EBITDA margin business.
Speaker #4: So essentially that's what we feel is the medium term and the long term.
Speaker #2: Oh, got it, got it, sir. And just if I can squeeze in one last question, in terms of guidance, any exit guidance in terms of MRR you want to give for this year?
Speaker #4: We, we don't, we don't give any guidance.
Speaker #2: Got it, got it.
Speaker #4: We always say that, like, look back at what we have done rather than giving a future guidance.
Speaker #2: Got it, sir. Got it. Thank you. That's it from my side.
Speaker #3: Thank you, Mr. Gulati.
Speaker #4: Thanks, Bharat.
Speaker #3: We have the next question on the line of Keshav from Niveshai. Please go ahead.
Speaker #2: Yeah, am I audited now?
Speaker #3: Yes, we can hear you, Keshav.
Speaker #2: Yeah. So, like, when should we expect the increase in GPU pricing to show up in the revenue? I mean, from the past six months, we have seen globally that the GPUs prices have been increased.
Speaker #2: So, like, when should we expect that to be reflected in the MRR? I mean, when the contract gets renewed?
Speaker #4: Well, see, as a company, obviously we have multiple goals. Which have to draw a balance between like, basically like, quarter on quarter performance and also our long term.
Speaker #4: Now, from the point of view of long term, like, we want to obviously have, like, a very good relationship with all our customers. And which also means that, like, we honor our commitment to our customers whether contracted or whether verbal.
Speaker #4: So in terms of, like, what they expect from us, and second part is that, like, we are heavily investing into our platform. Now, on our platform, of course, like, the increase in talent density both on the engineering and non-engineering side helps.
Speaker #4: And apart from that, like, obviously a big part of, like, build-out of our platform also requires that, like, we be able to use our own GPUs for our own product development.
Speaker #4: And not just be, like, a plain seller of all the capacity that we have. So obviously we want to heavily invest into R&D for helping in build-out of our platform.
Speaker #4: And any kind of R&D effort today obviously involves, like, necessarily using GPUs. So that's also an investment into the future. Now, that being said, like, again, like, we don't want to ever give, like, a future guidance, even if it is whether it is far into the future or whether it is for the next quarter.
Speaker #4: But as we kind of, like, go forward, like, obviously for uncontracted capacity or short-term contracted capacity, we obviously expect the earnings to grow from the same assets that we have.
Speaker #2: Got it. And one just last question for this question. Like, are you seeing the, you know, inference demand to be growing on platform and, you know, maybe it started contributing to E2E revenue?
Speaker #4: Sorry, I didn't question, like, you're talking about what's the contribution of, like, inference?
Speaker #2: Inference, inference. Yeah.
Speaker #4: Well, see, we look at the whole set of services as a overall whole, rather than, like, looking functionally at, like, over revenue is coming from inference or training or storage or CPU running agent.
Speaker #4: So we would continue to look at, like, the overall picture as opposed to kind of, like, breaking down things into numbers. Because quarter on quarter, all the numbers are subject to, like, potentially massive changes to any percentage we quote, like, we want to make sure that, like, that is something which is actually trackable and which is not, like, very ephemeral point in time.
Speaker #4: That doesn't, like, really help in terms of building and improved analysis around E2E.
Speaker #2: Got it. Thank you. I'll come back a bit.
Speaker #4: Yes, thanks. Thanks.
Speaker #3: Thank you. We have the next question on the line of Harsha from Silver Seven Rivers Holding. Please go ahead.
Speaker #4: Yeah, hi, hi, Harsha, please, please go ahead.
Speaker #5: Hi, hi, Tarunji. So, my first question is on the second tranche of V200. I'm just going through the balance sheet and couldn't spot any advances that we might have paid.
Speaker #5: So is there any delay out there? And I want to understand, what would be the rate at which these would be deployed? I believe the deployment is already committed.
Speaker #5: So would it be at similar rate at which we have deployed the 1,000 crore first tranche V100, or would it be higher or lower?
Speaker #5: If you can, you know, give some indication here.
Speaker #4: See, certainly our goal would be to make somewhat higher amount of money but, like I said, that, like, see, the long-term strategy of the company is not dependent on, like, oh, can we make, like, a few cents more or a few percentage points more from same set of GPUs.
Speaker #4: The long-term strategy is to kind of, like, make money by building more technology. And then the resulting growth is, like, a result of our advances in technology and customer adoption of our technology.
Speaker #4: So like, we start looking at it from the other side, where we say that, like, the asset growth is a result of kind of, like, growth of success of our technology rather than the other way around.
Speaker #4: I, I, I hope I am able to point you in the right direction over here.
Speaker #5: Yeah, yeah, to an extent.
Speaker #4: Like, essentially what I'm saying is that, like, let's not simply look at, like, the numbers, oh, in terms of, like, looking at GPUs as, like, a car-like asset, like, oh, can you make, like, somewhat more rental from the GPUs.
Speaker #4: Like, essentially GPUs are a means to an end. Essentially the idea over here is to get like, one, of course, the domestic market to adopt our technology and second, of course, to get our technology adopted, like, at a global scale.
Speaker #4: So the idea is not to kind of, like, run after short-term high hourly numbers on the GPU. Like, the numbers, like, would be a result of, like, when our technology gets more adoption.
Speaker #4: And not the other way around.
Speaker #5: Sure, sure. And selling Q2, how much revenue came from India AI mission? And I mean, are we expecting any renewals from, from those contracts to be happening at that better spot rate in going forward?
Speaker #4: See, we continue to engage with India AI mission, where, like, essentially the goal is to kind of, like, build relationships with various kinds of customers.
Speaker #4: Who will potentially continue to work with us beyond the remit of India AI mission as well. And regarding, like, numbers in the previous quarter related to India AI mission, I guess, like, Nathan can give, like, a better approximate answer about, like, what sort of revenue we earned from India AI mission related services.
Speaker #2: Yeah, so for the quarter, the revenue from India AI customer is roughly around 13 to 15 percent.
Speaker #5: Okay. And last question for FY28 and FY29. Just a follow-up, just a follow-up to previous one. So for FY28 and FY29, how does the allotment visibility look for maybe next lot of V300s or Vera Rubin?
Speaker #4: See, hard to predict the future, right? Basically, like, we will take the future as it comes.
Speaker #5: Sure. I mean, I mean, can we, can we take 2,000 GPUs per annum as a base case for next, next three years?
Speaker #4: See, we never, like, make any forward-looking predictions. So essentially, let's wait for us to kind of, like, buy our plans and then inform, like, everyone at the same time.
Speaker #4: That, like, what's the plan going forward? With regard to the increased APEX and deployments, et cetera.
Speaker #5: Sure, sir, sure. Thank you.
Speaker #3: Thank you. We have the next question on the line of Bhavya Gandhi from Bajaj Alternative Investments. Please go ahead.
Speaker #6: Yeah, hi, hi. Hi, Tarunji. Am I audible? Hi, hi.
Speaker #4: Yes, you're audible.
Speaker #6: Yeah, yeah, yeah. First question is regarding what sort of debt levels you are comfortable with and anything on the 1,500 crore enabling resolution for the fundraise that you've taken.
Speaker #6: And the second question is regarding what is the international revenue mix and what is the contracted revenue as a percentage of overall revenue. Yeah, those were the two questions.
Speaker #4: See, I think, like, from a capital deployment perspective, like, obviously we'll keep everyone informed in case, like, there is any future equity raise. So regarding debt, like, basically, like, I think it's also a function of, like, our own ability to repay the debt more than anything else.
Speaker #4: So there is no mental limit of, like, what can be the debt. Of course, there are, like, always covenants from lenders with regard to, like, what you can do at a given point of time or not.
Speaker #4: Now, there was another question related to, can you repeat the third question that you had? Like, that was related to.
Speaker #6: Sure, sure. If you can. Yeah, yeah. Perfect. But what levels you'll be comfortable, as a business, something that we are very comfortable with debt as long as we have the ability to repay.
Speaker #6: So we have, like, a upper limit defined for our business that, okay, like, we can't go beyond this, on that, or anything like that.
Speaker #6: So as long as the front-end capital supports, we are comfortable with that. Right. I was asking you about the international revenue mix and how much is the contracted revenue in the overall revenue by, yeah, like, what?
Speaker #4: Nathan, answer the numbers-related questions.
Speaker #6: Yeah, Tarunji, one more question regarding the asset-light model, if at all you are evaluating and if you can highlight what is your view on that.
Speaker #6: Before we go to Nathan, yeah.
Speaker #4: See, like, basically, like, this is a AI, like, this is basically like massive amount of AI build-out. Now, in an AI build-out play, like, essentially we are not ever to, like, any, any kind of, like, capital deployment that can essentially help us advance our technology adoption forward.
Speaker #6: Sure, sure, sure. Okay. And if you can, Nathan can answer the balance question, yeah.
Speaker #4: So in terms of the international revenue, which range from roughly around 20 to 25 percent of the revenue for the quarter is coming from the international customer.
Speaker #6: And contracted revenue as a percentage of the overall revenue?
Speaker #4: I think mix of long-term versus a short-term. Currently, our size is small to sharing of point-in-time information. And these numbers would change dramatically and.
Speaker #4: And Nathan, that also applies to the international numbers. So like, all percentage numbers we are looking at in the previous quarter, like, for quarter-on-quarter numbers, like, essentially are not, like, really trackable metrics.
Speaker #4: Like, they are subject to, like, wild swings is what I would, like, assume.
Speaker #6: Got it. But just as a 100, if 100 is the total revenue, what would be the contracted revenue that you'll be targeting? Just to get a flavor of annuity business.
Speaker #4: You know, like, there is no specific target in our mind. I think, like, it's a, it's a, it's a more of a tactical question which needs to be answered in multiple steps, which we kind of, like, keep changing.
Speaker #4: So like, there are, like, just so many considerations over there. Like, so where you can't, like, really say that, yeah, like, whether, like, you will target, like, a particular percentage as a contracted revenue.
Speaker #4: So that, that's not a fixed number in our mind today.
Speaker #6: Got it. Thank you so much. And wish you all the best for the future.
Speaker #4: Yeah, thank you.
Speaker #6: Yeah.
Speaker #3: Thank you, Mr. Gandhi. We have the next question from the line of Tushar Dhonde from Sanghvi Family Office. Please go ahead.
Speaker #6: Hello. Hi, team. Congratulations on a good execution. Two questions, I guess. First question was answered partially. That was on the pricing. The other half of the second question, first question was on the utilization.
Speaker #6: What sort of utilization levels we are in, and what sort of utilizations we plan to target? That first. And second one was on the TIR platform that we have.
Speaker #6: I wanted to understand how is the customer engagement we are having in the TIR platform, and what sort of competitive intensity are we seeing in this particular space?
Speaker #6: Like, are the global cloud operators having similar offerings, and we compete with them? Or, or we have a, a good product offering which, which is easily adaptable to the customer?
Speaker #4: Okay. So that's a lot of questions. So I only remember the last one. So from a would like to thank that, like, basically, like, whatever we are building, over here in India is, like, world-class.
Speaker #4: And can be globally adopted. That would, like, I would like to think. Now, the proof of the pudding rising lies in eating. So basically, like, the platform obviously the success of the platform shows up over a period of time.
Speaker #4: Like, over the medium term. So very, very hard to kind of, like, say that, oh, this is, like, better than X, Y, or Z today.
Speaker #4: It's a, it's a moving target. Like, so as the field of AI advances very rapidly, like, we need to build up our own pace of, like, keeping up with the latest, latest technology.
Speaker #4: And keep implementing, like, whatever is the latest thinking in terms of, like, how our customers are going to use AI. And I think that is what leads to the eventual platform as a product success.
Speaker #4: So I think, like, the jury for that should be held till, like, I think, like, the medium term. That's what I would like to think.
Speaker #4: Again, from a utilization standpoint, like, I think it's a point-in-time number. Shouldn't really matter. Like, what really matters is that, like, at an aggregate level, what the realization on the entire set of assets rather than, like, slicing and dicing and looking at, like, okay, this asset produced X, this asset produced Y, and so on.
Speaker #4: So like, the sources of revenue within, like, an overall platform can wildly vary between one customer to another customer. So some customer will pay for X, some customer will value Y more.
Speaker #4: They would pay for Y. So essentially, like, what I believe is that, like, utilization while it is a good metric to track assets which can produce very consistent set of average revenue, in case of a technology product, I think the variation is too high to say that, like, utilization is the right metric.
Speaker #4: So somewhere we feel that, like, the better metric over there is the overall realization across the entire fleet. Rather than looking at, like, the utilization metrics in terms of either the hours GPU hours utilized across the fleet or what number of GPUs were, like, idle and so on.
Speaker #4: Like, I mean, sure, we are using, like, our GPUs for, like, a lot of what is not immediate revenue, requirements. But that was required for, like, building up our product.
Speaker #4: I, I, I hope I'm able to get this across.
Speaker #6: Yes, yes, yes.
Speaker #3: Thank you. We have our next question from the line of Hitesh from Bloom Research. Please go ahead.
Speaker #6: Hello.
Speaker #3: Yes, Hitesh, you may go ahead.
Speaker #6: Thank you for the opportunity. Sir, I have two questions. So sir, how do you see E2E evolving in the next 18 months in terms of product offering and service offerings?
Speaker #6: And second question is, are you seeing any demand from global system integrators who are looking after our products for inference because their clients must be demanding for enterprise AI transformation?
Speaker #6: They must, must be demanding the GPU services, right?
Speaker #4: Okay. So one way to answer that question is that, like, look, we are very small. So we do see, like, all sorts of customers.
Speaker #4: But, like, today it's not really possible to slice and dice the set of customers we have. Because again, this is, like, very, very point-in-time over a very small base of the GPUs that we have.
Speaker #4: So like, are we talking to GSIs? I'm pretty sure that, like, that's, like, a part of, like, our overall sales coverage where, like, there would be members of our team obviously speaking to GSIs.
Speaker #4: So what can we say about E2E in over 18 months? I think that's a long period of time. So basically, like, I think, like, the strategy is flexible in terms of that, like, keep building technology that people want to adopt.
Speaker #4: And keep pushing the envelope in terms of, like, efficiency of our technologies which lowers the overall cost of ownership for our end customers. And kind of, like, keep building things that people value.
Speaker #4: So that's the overall strategy over there. And keep having, like, a great team that can kind of, like, implement the strategy. So that's as simple as that.
Speaker #6: And may I just, just a follow-up? Sir, recently you have launched a couple of inference-based products. Like, Blaze Endpoint and Jarvis Lab also. So are you seeing any traction?
Speaker #4: I, I wouldn't say launched as much as, like, basically, like, released. Like, soft release. Like, obviously, the official launch might be, like, a few weeks or so away.
Speaker #4: So yeah, like, basically, like, that's the general direction we are taking. But, like, we are kind of, like, building a lot of cool things that our customers can adopt.
Speaker #4: And help us learn about, like, what they want to do with their tech that we are building. And as we start seeing substantial adoption that can show up in the volume, we'll start talking more about it.
Speaker #6: Okay, sir. Thank you.
Speaker #3: Thank you. We have the next question from the line of Neil Manath from Pico Capital. Please go ahead.
Speaker #4: Yeah, hi. Hi, Neil.
Speaker #6: Hi, sir.
Speaker #5: Hello.
Speaker #4: Yeah, please go ahead.
Speaker #5: Hello.
Speaker #6: Yeah, am I audible?
Speaker #4: Yes, yes, you're audible. Please go ahead.
Speaker #6: Yes. Hi, sir. So my first question is with respect to the current Indian competitive landscape. I mean, we've got one player that has scaled up to 90,000 GPUs, and we've got another very recent entrant that has got, I think, 8,000 odd P300s.
Speaker #6: So, I mean, and this is related to our statements with respect to our platform. So, I mean, how do we differentiate from these players when they have already scaled up to these levels?
Speaker #6: I mean, when do we see a similar kind of scaling up? Since we keep pressing about our platform in that sense.
Speaker #4: Sure, Neil. So let's take a step back. So I think, like, the overall total addressable market for AI infrastructure alone at a global level, like, we are literally talking about, like, over, like, the medium term, like, maybe anywhere close to, like, a trillion dollars of AI infrastructure addressable market.
Speaker #4: Now, AI is today, like, a truly global business. So the GPUs are facing wherever there is power. And wherever there is, like, availability of, like, ready-to-release, like, what do you call, land cover shell, so essentially, what we are looking at today is, like, a global marketplace.
Speaker #4: And the global marketplace, like, I think, like, millions of GPUs are getting deployed, like, every year. So when we talk about the entire set of GPUs being deployed in India, like, it doesn't change the fundamental equation for us in terms of what we are doing.
Speaker #4: So there is no what do you call, like, direct competition between, like, basically, like, any two, three, four, five players which could be there and, like, one particular market segment.
Speaker #4: Essentially, we are looking at, like, a very large-scale global market where everyone who is a player in this market needs to find their own niche.
Speaker #4: They need to find their own set of customers where they learn from those customers and then double down on their learning to build technology that can be used by customers similar to the customers they have.
Speaker #4: Or maybe even expand into, like, newer customer bases or industries. So essentially, in a long-term AI supercycle build-out, like, all of us will need to figure out, like, what we want to do.
Speaker #4: Now, we have a reasonably mapped out agile strategy around building technology. And getting the adoption and learning from our customers and working with them to kind of, like, give them what they need and we hope to be able to continue to do that.
Speaker #4: Like, so there are there is only, like, limited set of people solving problems in AI, like, even globally. And of course, more and more people are coming in and solving more and more problems.
Speaker #4: But it's still leaves, like, a significant set of problems to be solved by us, like, which can be adopted by our customers. And I suppose that, like, that will give us the ability to continue to grow, like, year after year or at least for the medium to long term.
Speaker #4: So that's, that's, that's what we think about, like, basically the competitive intensity in this space. Like, so we are in the midst of one of the largest build-outs, like, there are more than, like, 150 or 200 or 250 odd players in the world who are pursuing this market.
Speaker #4: I think there is enough for everyone.
Speaker #6: So, sir, I'll just rephrase my question here. I mean, in a similar context, when we are serving Indian customers, and when Indian players serving Indian customers, if price is same, and they are running ahead of us in terms of addressing demand, I mean, how then, then our platform differentiates from their platform?
Speaker #4: Okay, like, I said, like, we don't look at this as, like, a Indian term alone. We are looking at, like, us as a global player operating in a global market.
Speaker #4: So essentially, there are enough and more customers for everyone in the market today. So I think, like, this may be a question for, like, far into the future that, okay, like, there is not enough demand.
Speaker #4: How are you faring? How are you going to kind of, like, try and beat your competitors? So like, I think, like, for next couple of years, like, probably this is not even a question.
Speaker #4: Like, so like, everyone is trying to figure out AI. Everyone is trying to build technology on AI. Everyone wants their technology to be adopted by others.
Speaker #4: So like, we are one amongst, like, many, many, many players in AI. We are trying to push the envelope. So and, and the market is large enough.
Speaker #4: Like, any, any particular niche you take in a trillion-dollar market is, like, a niche which is worth, like, maybe 10 or 20 or maybe 50 billion dollars.
Speaker #4: So I think, like, we feel very, very comfortable in this market. Like, it's a vast blue ocean. There is no red ocean today that, like, we need to worry about competition today.
Speaker #4: So we are very comfortable.
Speaker #6: Just a follow-up on the sorry to interrupt. He's fine. You go ahead. Just a follow-up on the previous participant's question. With respect to, sir, are we exploring off-balance sheet funding for any of our GPU capex?
Speaker #4: See, essentially, we are we continue to operate in an ecosystem that exists today. And in this ecosystem, like, everything exists. And like, we have a saying, never say never.
Speaker #4: So we are super happy to explore any and everything with regard to, like, what we are doing on tech, what we are doing in terms of, like, funding, for the GPUs, what we are doing in terms of distribution.
Speaker #4: So like, we are open to everything to kind of, like, keep pushing our strategy of, like, getting our technology adoption over a period of time.
Speaker #6: Okay. And if I can just squeeze in one more question.
Speaker #2: Sorry to interrupt, Sunil. We have other members also in line. The question kindly come up with a follow-up question in the queue. We have the next question from the line of Pratik from Bandon.
Speaker #2: Please go ahead.
Speaker #6: Hi, sir.
Speaker #4: Hi. Hi, Pratik. Yeah.
Speaker #6: Hi. Hi. Sir, I just wanted to check. See, we have two platforms, right? Public platforms, the Thir and Jarvis Labs. Is there so at the back end, are the engineering capabilities shared, or, or there is duplication over there?
Speaker #6: And the second is, obviously, you know, you will move from developer, workloads to production workloads, et cetera. So how does the handover happen between Jarvis to Thir or, or at the front end, is there separate go-to-market strategies?
Speaker #6: How does that work?
Speaker #4: See, both are opinionated platforms on solving problems for their own set of customers. There is obviously going to be overlap in terms of some of the tech that gets built, on both the platforms.
Speaker #4: And there is going to be, like, some overlap in terms of customers as well. But eventually, like, both the teams are sufficiently different from each other in terms of, like, where they want to kind of, like, take the final technology product.
Speaker #4: And the final technology journey, I wouldn't even say, like, it's a somewhere that journey is stopping. So essentially, it's a journey to undertake by both the teams.
Speaker #4: And like, both have so there is enough and more to build in, like, a massive 1 trillion dollar term market. So I think, like, there is no upper limit to the amount of technology you can build.
Speaker #6: Understood. And how do you ensure that both don't compete at the front end in terms of pricing, right?
Speaker #4: See, I just answered this question. This is, like, a multi-million GPUs per year kind of a marketplace. Like, I think there is there is enough and more for, like, everyone not to step on each other's toes.
Speaker #6: Okay. Fair, sir. Fair. Thanks. Thanks. This is the only question I had. Thanks, sir. Thank you.
Speaker #2: Thank you. We have the next question from the line of Hardik Goyal. From Union MF, please go ahead.
Speaker #4: Yeah. Hi. Hi, Hardik ji.
Speaker #6: Hi, sir. Hi, Tarun. Thanks for this opportunity. And congrats on a good set of results. Sir, you have you are not commenting on the utilization, but the 1,000 crore contract that we got, for our B200, does that completely exhaust our capacity of the current B200, or do we still have some ideal capacity on that end?
Speaker #6: And secondly, sir, how is the current supply chain scenario on the GPU procurement, and how are GPU players compared to the peers in that scenario?
Speaker #6: Thank you.
Speaker #4: So see, like, again, like, I'll start from the last question. Like, so there is no peer comparison. Like, so we have our own tech journey to undertake.
Speaker #4: So we are not focused on, like, basically, like, any peers in that sense. Now, from a very specific question in terms of, like, whether the current maximum capacity is 100 percent utilized, the flagship maximum capacity is fully utilized in the particular contract that was mentioned.
Speaker #4: And yeah, was there a third question? Sorry.
Speaker #6: No, sir. Only this. Sir, on this, are we still facing some backlog in the orders from NVIDIA?
Speaker #4: See, like, it's a it's an organic cycle, basically. Like, you plan ahead, like, and you project your basically, like, very agilely you kind of, like, react to the market and predict for the near future, like, what you want to deploy.
Speaker #4: And as those plans rectify, like, obviously, like, we would keep everyone informed. So like, I don't see, like, any fundamental supply chain challenges that are unsurmountable for us, like, in the near or medium term.
Speaker #6: Okay. Sir, last question. There is a new GPU launched by AMD. So is our system compatible to use that GPU in the future? Is it a case or situation?
Speaker #4: See, for me, software point of view, like, obviously, the software that we are building, we would like it to be compatible with anything that is out there.
Speaker #4: But more specifically, for that to happen, like, obviously, we'll have to engage with, like, any and every chip maker, which usually we do. Like, so in the sense that, like, there are a lot of people building chips these days, and we obviously like to keep our eyes and ears open to what is happening.
Speaker #4: And engage technically to understand, like, everything that is out there. And obviously, if there is, like, support requirement from our customers for our software to support something, then, like, we are obviously very, very willing and able to build that support.
Speaker #6: Perfect. That answers my question. Thank you, sir. And all the best.
Speaker #4: Yeah. Thank you.
Speaker #2: Thank you. We have the next question from the line of Sonal Sharma from Kapra Investments. Please go ahead.
Speaker #4: Yeah. Hi. Hi, Sonal ji. Yeah.
Speaker #2: Mr. Sonal, can you hear us?
Speaker #5: Hello. Hello.
Speaker #2: Yes, sir. We can hear you.
Speaker #5: Hello. Yeah. First of all, congratulations, sir. For the exceptional results. My question was on the growth story. So on the deployment part, you had mentioned 1,024 plus B200 planned GPUs, right?
Speaker #4: Right.
Speaker #5: So like, can you know the timeline and the funding that you guys are funding?
Speaker #4: Just what are we expect the deployment to go through?
Speaker #5: Sorry?
Speaker #4: This, this quarter we expect the deployment to go through for the 1,024 GPUs.
Speaker #5: Okay. Okay. Okay. And any feasibility on B300 or better looking?
Speaker #4: Like, so we will obviously kind of, like, keep everyone informed as our plans for Vera Rubin and B300 rectify.
Speaker #5: Okay. Okay. Okay. Thank you.
Speaker #4: Yeah.
Speaker #2: Thank you. We have the next question from the line of Nikhil C from Toro Wealth Managers. Please go ahead.
Speaker #3: Yeah. Hi. Hi, Tarun. Good evening. And congratulations on a good set of numbers. I'd just two questions. Since we keep mentioning our long-term strategy is to sell GPU as a part of technology platform rather than just raw compute, I just want to understand how does probably our economics change when probably the transition happens with a customer maybe a live customer.
Speaker #3: Like, how does the realization change? How does the gross margins change? And probably the ROCs on the GPU capex change?
Speaker #4: See, I think that's a question that would get answered over a period of time. So I don't think there is, like, a answer to that today.
Speaker #4: But like, essentially, if you look at, like, any kind of technology play, like, so when you build layers of technology, and you get the adoption from the customers, then obviously, long-term value gets unlocked for those customers who adopt that technology by reducing their own TPO or building new capabilities for them that allows them to move faster.
Speaker #4: And that essentially results in, like, surpluses, like, up and down the ecosystem. So like, hard to put, like, say, numbers to that today. But like, that's where the adoption of tech is, like, a medium to long-term strategy.
Speaker #4: So that's, that's where we are today.
Speaker #3: Understood. Understood. So is it fair to understand, as we speak currently, probably most of our revenues is primarily the GPU raw compute? Like, what we are giving it to our customers?
Speaker #4: Minority of our GPUs, like, today run on top of our platform. So like, there is very, very little bare metal that we sell today.
Speaker #3: Understood. Understood. Thank you so much. I'll come back in a bit.
Speaker #4: Sure.
Speaker #2: We have the next question on the line of Varun Gandhi from FinAvenue Growth. Please go ahead.
Speaker #4: Yeah. Hi. Hi, Varun ji.
Speaker #6: Hi, Tarun. Good to catch up again. Tarun, I noticed you guys attended a lot of conferences, a lot of trade shows, this quarter, and these two quarters particularly.
Speaker #6: Anything fruitful that you could share, any client feedback, any conversation that helps you improve your product, helps you gain some sense of, you know, from a technology standpoint, how competitive are you?
Speaker #6: Anything that you could share?
Speaker #4: Okay. So see, essentially, like, it's a, I don't know, like, there are many worlds. So basically, I think, like, even while we were, like, in the heart of Silicon Valley attending the biggest AI conference organized by NVIDIA, I think, like, a couple of blocks away, like, the conversation was not about AI at all.
Speaker #4: So
Speaker #7: from late
Speaker #8: any one place with one set of people working on tech versus another set of people working on completely different say sectors, like there is a lot of what you call gap in terms of like how people perceive and look at the how they are going to deploy and how they are going to use the technology.
Speaker #8: So I think like the magic happens when you are able to bridge the gap between say you're thinking of a customer in a particular sector with how you can apply the technology that you know.
Speaker #8: So technology does not like really solve the problem. It's the people who have to solve the problem. So technology is more like, okay, we are going to run a training.
Speaker #8: Now, does training solve like a real-world problem? Probably no. For someone in a fintech or a finance sector or insurance sector or a molecular research sector, like you have to bridge the gap in terms of like what AI can do as a technology, and how do you bridge the gap between the objectives of a customer who wants to deploy that technology and kind of like do stuff.
Speaker #8: So that's where I think like you bridge the gap with the help of industry solutions where you say that like, look, we know this part very well.
Speaker #8: And you know this part very well. How can we collaborate together and co-design a solution that works for you? And that also helps us learn something.
Speaker #8: About a particular domain and like how technology can be applied to that particular domain. And I think it's the constant process of being engaged in the ecosystem where both the hands-on technologists and decision makers frequent.
Speaker #8: So that's the overall idea behind like being
Speaker #9: तो सरों हैज़ हैज़ एनी ऑफ़ इट फ्लोन हैज़ एनी ऑफ़ इट फ्लोन इनटू द बिजनेस
Speaker #8: It's a decision process. So like any, any point in time learning are like exactly what they are, like point in time learning. And as and when we apply those learnings, like obviously we try and publish those case studies.
Speaker #8: So, but that's the look back rather than saying that, oh, we are predicting that this is what is going to work. So I think that's a longer journey than you could predict, like, but just by talking to people.
Speaker #8: So yes, of course, there are very many experts who know better than us in terms of like how they see, how they perceive the future.
Speaker #8: But then I think like somewhere in a very large world, in a very large town, you have to be having the humility of like saying that like we don't know like what is going to work, but be agile enough to say that like yeah, like let's make it work together, like whatever is going to work.
Speaker #9: Got you. My second question is related to L&T. So they recently acquired 10,000 B300s, and I understand you cannot comment on L&T and their business strategy, but then we as partners could use their idle capacity if they have, or spin up some sort of partnership whereby you use your platform and the capacity provided by— the GP capacity provided by L&T.
Speaker #9: Any sense, any ideation, anything there?
Speaker #8: We have a completely arm's-length relationship with L&T where like they are a valued partner, like where we work together with them on, uh, for servicing many of the customers.
Speaker #8: Uh, that being said, that similar type of relationship we are building with many other partners as well. So essentially, we are always open to all possibilities of working with all our partners.
Speaker #8: Now, these things happen when they happen, and like when they happen, we'll obviously keep everyone informed. But like in the ecosystem, like there are very many players and like basically all possibilities of working together with all the players.
Speaker #8: continue to remain open.
Speaker #9: So it's safe to conclude that you guys are open to such an arrangement, and if, if possible, you might get like everyone in the ecosystem.
Speaker #9: Understood. That answers my question. Thank you very much. Great, great.
Speaker #8: Thank you. Ladies and gentlemen, we will now be taking the last question in the interest of time. I now invite our next speaker, Kaushal Sharma from Equinox Capital.
Speaker #8: Please go ahead.
Speaker #9: Hi sir, very good. Yes, you're audible. Please go ahead. Yes, congratulations on good set of numbers. Just want to understand your procurement mechanism of the GPUs, like how many months or a period do we require to place any advance order for the GPUs to get it went live.
Speaker #9: And sir, if— do we require to pay any advance against that? And if yes, so what percentage of CapEx is it? All these answers
Speaker #8: are like subject to change based on market conditions. So like, uh, like we find that like this, uh, uh, order to deployment cycle can vary between 3 to 6 months.
Speaker #8: Typically. And like whether an advance needs to be paid depends on like the vendor, so whether they have a requirement of like collecting the advance at that point of time.
Speaker #8: So probably depends on their order book in some way. So like everything is like kind of like subject to massive changes like week on week or month on month.
Speaker #8: About all the questions that you have asked.
Speaker #9: But currently, if there is too much shortage of the GPUs and new, new GPUs are coming, so don't you think that we need to pay any advances of current scenario demand and supply?
Speaker #8: Like, I am not sure how to answer that.
Speaker #9: Okay, sir, got it. Thank you very much for answering question. Thank you.
Speaker #8: The next question from the line of Bharat Gulati from Dalal Ambrosia. Please go ahead. Hi, hi, hi, thank you for the follow-up. Thank you for the follow-up, really appreciate it.
Speaker #8: Uh, just on the margins again, so just I understand that you spoke about maintaining 70% plus EBITDA. But in terms of what our revenues are right now and what gross margins we've done for this quarter, in the medium term, is this gross margin sustainable of 88%?
Speaker #8: And if you can just help us understand what has helped us achieve this 88% gross margin in this quarter also. I see, it's a— it's all of these numbers are like moving targets.
Speaker #8: So essentially, like, uh, like, uh, these are like not like really quarter-on-quarter target that like basically like you can, uh, so I'm pretty sure that like if in a particular quarter we do like a lot of CapEx at an inopportune time, like, uh, so how these numbers change at that point of time, or like if you build some capacity in advance of and it's not yet 9.
Speaker #8: So I guess like so many factors are there. But Tarun sir, would it be fair that at full utilization this gross margin is what we have been saying that like look this is like a 70% plus kind of an EBITDA kind of a business.
Speaker #8: So I don't see any change in that overall target for the business. Understood. And so just on the next Blackwell that we are planning to install, do we already have any contracts set for that, or are we planning to get into a long-term contract for that as we've done for this batch?
Speaker #8: Just trying to understand how do we We have not decided anything around the next lot of B200. Understood. Thank you. Really appreciate taking the question again.
Speaker #8: Yeah, thanks. Thank you. Ladies and gentlemen, due to time constraints, that was the last question for today. We understand that several questions may have remained unanswered, and we encourage participants to reach out to the company through its investor communication channels.
Speaker #8: Including email for any further queries. I now hand the conference over to Mr. Tarun Dua for his closing comments. Okay, uh, first of all, again, I would like to thank everyone who has patiently listened, joined our call, and patiently listened to us talk about like how we think.
Speaker #8: And we do welcome all investor queries to either through our investor relations team at Airtector Or you could reach out to us via the email ID investors@e2enetworks.com, and we are super happy to engage with all of you to try and build a better understanding of like how we feel and how we think about this business.
Speaker #8: And we hope that you have a great festival season including Diwali season coming up ahead. And once again, thank you, all of you, for joining us here today.
Speaker #8: Uh, good night all. Thank you. On behalf of E2E Networks Limited, that concludes this conference. Thank you for joining us, and you may now disconnect your lines.
