Participating in this session of the WTR Insights Conference featuring Gorilla Technology. My name is John Roy. I am Managing Director for Technology here at Water Tower Research. We are pleased to be hosting Jay Chandan, he is the CEO, and Bruce Bower, the CFO today. Welcome, gentlemen.
Thank you very much, John.
Thanks for having us.
Now, before proceeding, I must note that Gorilla's safe harbor statements are available on their website. As always, investor questions are encouraged. Please enter them into the chat box, and we will address them as soon as possible within the session or with a forthcoming management series report. Those wishing to request a meeting with Gorilla Technology can do so through the conference portal. With that, let us go ahead and get started. The first question of the day. Really, what are the primary revenue drivers that are really powering Gorilla Technology's Edge AI, smart cities, and cybersecurity, et cetera?
John, once again, thanks for having us. I think personally, the numbers tell a better story than I can, right? We did $74 million+ of revenue in 2024. We said we were going to do about 100. We did $101 million in 2025. I think that growth is roughly around 35%, 36%, 37%. The first half of 2026 alone, we have done a little over, what, $78 million+ of revenue. That was almost 100% growth year-on-year. Now, what is very important is that it shows Gorilla is a different business from where it was when it started out. Even 18 months ago, we were a completely different revenue generating business, and there were three revenue engines. Now, the first most important, it is an established business, and I want people to understand this. We are a software house.
Edge AI, security intelligence or security convergence, as we call it, smart cities, network intelligence, cybersecurity, that's our bread and butter. These were all primarily what I call government-based contracts, long-term, public sector, large enterprise deployments, and so on and so forth. They remain an important foundation of this company. The second, which has grown extensively much larger, which is a growth engine today, is the AI infrastructure piece. This is what we call the infrastructure, the GPU as a service, and the associated AI infrastructure, which we are building over the next few months. Yotta is an obvious example for that. Our expanded program in India covers roughly around, what, 20,736 Nvidia B300 GPUs, and has an aggregate contract value of roughly around, what, $2.5 billion-$2.6 billion. That alone is the beginning.
If you look at the other 640 servers, which is building roughly around 5,000 GPUs, give or take, you're talking about another additional $500 million+ of revenue. So in India alone, we're looking at about between $3 billion and $3.1 billion of revenue. The third engine, which is all about infrastructure, is around data centers. It's around managed infrastructure. It's about sovereign AI and ultimately the software and cybersecurity layer sitting on top of that compute. We've been talking and trying to educate the market over and over again. What is exciting for me is the convergence of all that. We used to sell intelligence layer. Now we are increasingly participating in the compute, in the infrastructure, in the software security, in the security convergence layer, and the managed service economics.
If you put all this together, that dramatically increases our revenue per customer, and that's what we're trying to do. Our revenue per customer increases so that we can do all the things right and nothing wrong before breakfast. That's the business we have chosen, and that's where we're going to today.
Excellent. If we can go for a specific extremely large project, Yotta AI infrastructure. Maybe you can give us some color, where you're at on that, what milestones are you looking at, et cetera. Just give us everything you can.
Absolutely. I think, Bruce, I am happy for you to chime in. We kind of gave evidence to the market that we were already deploying. Yotta has moved from being a very exciting contract to being an enormous execution exercise. If people have been living under a rock, they wouldn't have followed what we are in terms of the hiring, but we now have doubled our hiring already. We are now including the contractors. We are plus 600+ people already. The first program consists of, like I said, 640 or 5,000 B200 GPUs. The second phase consists of 20,736 B300 GPUs. We have delivered the first phase. There are four phases for the first of Yotta 1. We have delivered the first phase.
The remaining two phases are being delivered over the next week or two as we speak today, and the final phase will be delivered before the first week of October. Max second week. The Yotta 2, which is the 20,736 GPU, B300 GPUs, they will start being delivered by the first week of October, in blocks of 5,000 GPUs. You are looking at roughly four, if not maximum five delivery schedules, and they all have to be completed before the end of this year. If you are looking at a collective infrastructure, we are looking at about 25,736+ GPUs being deployed. Actually, that is a wrong number. It is actually 25,856 advanced GPUs because it is 5,110 B200 GPUs. If you take the execution, Supermicro has already been working very hard with us. As you know, we have a very, very strong relationship with Supermicro.
We also have managed a supply chain arrangement with them, and I think that is approximately, roughly about $2 billion+. We are no longer on PowerPoint numbers anymore, John. The real procurement, the commitments, the deployment schedules, the data center capacity, the networking equipments and so on and so forth, all seem to be based on the customer delivery milestones we have committed. That is why I keep telling this organization, the only word we choose to use today is execution. Every single facet of the organization, every single silo we have built, is focused on execution. We are applying the same discipline to all of our programs today.
Bruce, you are okay with the funding for this?
Well, I wanted to add, the important thing is that this equipment gets delivered, and then it's a few weeks for us to set it up, configure it. Then it goes to the client for acceptance and testing, and then it's billing. When people talk about the revenue forecasts, when we have given guidance, it's based on the understanding that this equipment is deployed and then it's billing, and so it's billing by the end of the year. So for Yotta 1 and then Yotta 2, that is where the guidance is rooted, is basically these are the delivery schedules. It gets deployed, it gets turned on, and then it gets billing. That's how we get to $450 million or $500 million as our preliminary guidance for next year.
Excellent. Now, obviously you guys have been doing a lot in Asia for many decades. Maybe you can talk about how you're positioning yourself to capture really the next phase of the AI infrastructure and sovereign AI out in Asia. It seems to be Southeast Asia in particular is a hotspot for you. Can you give us some color and some details as to where that's going in that market?
Absolutely. Before people kind of pigeonhole us as Southeast Asia, most people understand we've actually built a whole bunch of data centers in Egypt as well, and that's for the government of Egypt. You can see that they're all running successfully. I don't know if most people have realized this, John. I think about a month ago, the president launched the Octagon. The president of Egypt launched the Octagon, and our infrastructure sits on top of it. I cannot think of a better way to actually give people an example as to what Gorilla has actually done and built for a very, very significant government like Egypt. Coming back to your question, we've actually been focused on South Asia, Southeast Asia, and Far East Asia, and we've kind of divided them into three different buckets.
In South Asia, obviously you've seen the whole program with India, with Yotta. We're already kind of, like I said, 25,000+ GPUs already being delivered. We're also in active engagement conversations with Yotta on further deployments within the Indian subcontinent. Look at Indonesia. We've already gone and signed up for, what, 1,875 servers, and that approximately is about 15,000 GPUs. That has to be deployed by the first quarter of next year, as you've already said. Fast-forward, we've also taken into account our Korat AI data center campus. Now people are going to be asking, "Hey, what's happening with that?" The government's already allocated 280 MW across, and we are building four data halls across four different phases. We're also developing activities across Malaysia. We've also signed additional capacity in Indonesia between 135 to 200 MW.
It's no longer becoming a Southeast Asia story. It's becoming an Asia story, and we're kind of honing it down to make it an ecosystem for us. India gives us enormous what I call sovereign and enterprise AI demand. Indonesia gives us a very large compute market. Thailand becomes what is called a strategic infrastructure hub for us. Singapore remains extremely important for capital management, regional structuring. More importantly, if you look at Malaysia, it's growing as a fast-growing data center market region as well. Along this, we are also exploring the U.S. We're working very closely with a couple of opportunities in the United States. We're exploring opportunities in South Korea, in Japan and so on and so forth.
Remember, I think we mentioned this in our, or not I think, we mentioned this in our previous earnings release, our guidance conference call, that our company's pipeline is north of $7 billion, including the AI infrastructure, the data centers, public safety, and other opportunities. What people don't realize is that we've actually not just expanded that pipeline. That pipeline is well north of $10 billion+ right now. Not all pipeline becomes revenue immediately. I just want to be very clear, but I want to make sure that everybody understand. It demonstrates scale. We're not talking about building a pipeline of $100 billion, unlike other companies in the market. All we're saying is we have a very solid pipeline of $10 billion+, which we will execute over the next six to eight months as well. The really interesting part is where are we going next?
For me, the next three to five years is predominantly focused on sovereign AI. Countries increasingly with the view, they've kind of changed their strategy. They've gone from compute capacity to saying, "Look, I want my entire energy infrastructure here, but I also want all of my national AI capability not sitting somewhere else in the world, but within my country." That is creating an extraordinary opportunity for us to actually combining compute data centers, all our networking, cybersecurity experience, and sovereign architecture experience to bring together. That's why Gorilla is completely positioned. I keep joking about this within the company. I said I wanted to run a software company, and somehow I decided to build half of the infrastructure for Asia. So I'm quite happy with that. But it's daunting and, don't get me wrong, it's really hard work.
These projects you're doing, they're not build it and then look for a business. They're bespoke, aren't they?
Yes. They are absolutely bespoke. Yes.
All right. I just want to be clear, because a lot of people say, "Oh, they are just building data centers." And it is like, yeah, not quite.
I think.
John , to be clear, we do not move a finger without a customer.
Right.
It is a great point.
Exactly. Good point. Bruce, I know the recent results, there were some adjustments. I really wanted to get your view on what margins are going to do over the next period here, and how you're going to manage cash. Certainly these projects are not small. They take a certain amount of effort, and all the investors want to make sure you're just not buying business, right?
Yeah.
Yes. The margins were a little bit weaker than we would like on the gross margin and the EBITDA margin in the first half of 2026. That was due to two things. One is we had some foreign exchange fluctuations, which pushed down the gross margin. We paid earlier and then the exchange rate depreciated. By the time we deliver the solution and recognize the revenue, there's a mismatch, and that drags down margins. The second thing is that we took a couple of large one-off expenses, and that's what hurt the operating and the EBITDA margins in the first half. What I can say is that the large one-off expenses won't recur. By definition, they're one-off. Going forward, we see a program of normalization of margins. Back to the sort of traditional Gorilla margins and then expansion.
The normalization is driven by the fact that some of these FX movements, assuming a normal FX environment, they won't be there. The pipeline, as Jay Chandan talked about, is really in the security and the traditional business for Gorilla at sort of 30%+ margins on a gross margin level. In the GPU-as-a-service AI data center market, that's 70%+ gross margins. As we mentioned, some of these projects will start billing before the end of the year. As we add next year, $300 million+ of new data center revenue, at least, then that's at 70% gross margin. You'll see the overall gross margin for the business go up and also the EBITDA margins go up. I would say that it's safe to model a 50%+ gross margin for the entire business next year.
In terms of EBITDA margin, it's early to say because obviously we're in expansion mode with hiring many people and then also some other costs, but I would say 20%+ EBITDA margins is something we can shoot for.
Yeah. Actually, if I may add to that as well, Bruce, I think one of the positives we actually saw is that we've been collecting money from our customers. Like Egypt, we announced in the beginning of the year money was being collected, solutions being delivered. We've been invoicing the customer. There's more money coming in in the next few weeks. More importantly, we've also been winning. These are not just data center projects. We've also been winning large deals. I think the market also needs to understand that we haven't forgone, or we're not trying to forgo our revenues coming from cybersecurity. For example, we recently won a 5G LI project, again, with the government of Taiwan. Guess what? That pays the bills, right? Those are high-margin businesses. These are multi-year contracts and in tens of millions of dollars.
They go from year to year to year. So I think people need to understand it's not a data center infrastructure story. It's an AI infrastructure story. That's why I keep telling the market, AI infrastructure. We do 5G LI, we do cybersecurity, we do networking. Guess what? Unfortunately or fortunately, the AI infrastructure part of it, the data center part of it, is just a bigger chunk of revenue. That's about it.
Then, John, the second part of your question was about how we manage cash. So Jay, there are several platforms or several things that are important to keep in mind. The first is what Jay just mentioned is cash discipline within the business. So the collections have been strong this year. We expect the sort of traditional Gorilla business focused on cybersecurity to generate cash flow this year. It depends on the speed of collections, but it'll be in between sort of $5 million- $15 million of cash generated by that business. Then in terms of how we manage the cash going forward, we've been very clear in the past year about our strategy has been first we raise equity or quasi equity in the form of convertible notes.
That is our kind of contribution to projects, then we raise external financing, primarily in the form of project financing. That finances the payments for the equipment, the delivery, and then that is paid down over time by the revenues from the project. That remains the core of our strategy. We got a credit rating from Fitch, B- with a positive outlook, and that is helping with that effort, of course. Now they know exactly where to place us and how to price us. I can announce more later. It's confidential at the moment, but I think that the market will be quite happy with what we've done on that front.
Excellent. Well, maybe if you could just remind us what is your guidance for next year or your outlook, if you haven't given official guidance as yet.
Official guidance is $450 million- $500 million of top line.
Excellent.
John, our ambition for year after, just in case, ambition, okay? We're not putting out guidance yet. We're looking at anything between $1.3 billion- $1.8 billion. We're growing.
Let's talk about the overall market. The overall market, obviously there's many governments interested in AI. I was curious is how do you see you're going to be able to keep a competitive edge? In other words, let's be honest, why are you going to keep winning?
Why? That's a great question. We've been converting very large contracted programs, as you've already seen. The reason why people come to us is a multifold of reasons. First things first, we understand large-scale sovereign AI. We've been working in this business for 25 years. For people who don't realize this, our team collectively has built more than 40+ data centers to date. When we go into data center platform programs, we suddenly become quite relevant. More importantly, we also not just provide GPU as a service, we can actually provide our entire networking infrastructure. We build the SOC, the NOC, the BMS solutions, the entire tunneling, the network analyzer. We do all of the cybersecurity and so on and so forth.
Infrastructure is not just about acceptance ramps, your utilization ramps, and your deployment ramps, and putting some nice spreadsheets and so on and so forth. For us, it's about making sure that we are able to deliver. To the customer today, we are making sure that we're delivering on time. We talked about the business is not just about getting the GPUs and funding it. It's also making sure that we're able to get the financing at the right price. Bruce talked about revenue, right? You asked a question previously about revenue. Revenue is nice. Profit is even nicer for us, right? It's the cash that actually pays the bills, John.
Between Bruce and I, we have about 17 calls a day which says, "Hey, how are we on cash?" I think that's what makes us very different and dangerous as well, because we are not just keeping ahead of the business today. We know exactly where we have to be in the next three years. If I say five years, I'm going to be an idiot. But three years is definitely, I know exactly where. Whether it's building into the modular data center businesses, whether it's acquiring power plants, whether it's acquiring submarine sea cables, we are looking at it all. That means our entire infrastructure comes as one.
Excellent. Well, maybe one final question because you're running against the time limit here. What do you really think of the market's misperception, I guess? What are they missing about your story that's the most important?
Bruce, do you want to take that first and I can close it?
Yeah.
Yeah, I would say what they miss is that it's happening. Like you said, it's not just a PowerPoint presentation. We have announced projects, and then we've progressed them to execution phase. With Yotta 1, for instance, we've taken delivery of a batch. We're taking delivery of more batches of equipment. Yotta 2, which is the larger contract, should be executing by the end of the year. This isn't sort of, "Oh, I hope that those projects convert." This is a case of they are happening. There's the guidance, but if you were to sit down with pencil and paper and add up the annualized revenue from Yotta 1, Yotta 2, and then our project with NeutraDC, that would be about $800 million of annualized revenue.
It's happening, and if it all happens as we plan, then I think the targets are realistically well above our current guidance. The thing I would say is that what we're telling people is the story, right? We're trying to build credibility with the market in every direction. First, like I said, is credit rating and really building more credibility, deep relationships with the financers. That's private credit, that's banks, that's Wall Street. Then trying to get the story out and then helping research analysts who want to cover us. We've had coverage picked up by Zacks, coverage picked up by B. Riley, by Compass Point, and then we have a few more conversations going. We're out there, we're telling the story, and we're delivering.
That's a great point, Bruce, actually. That kind of goes into what I wanted to actually talk about as well. Institutional credibility ultimately for me comes down to numbers and how we behave as a company. We continue to deliver. I think market doesn't seem to understand this company has gone from $22 million of revenue, with an $85 million - EBITDA, to where we are today. The market doesn't give us credit for that. That's okay. We will continue to deliver because the market will eventually catch up. Second, we've increased our balance sheet, materially increased, strengthened our balance sheet. Right? If you look at the way our business has evolved, whether it's share purchases, and I don't know how the market seems to be missing this. We've done share purchases repeatedly. We've done buybacks.
Even up until February of 2026, we've been buying shares from the market. We moved from $5 million projects to $10 million projects to $100 million projects. Now to multi-billion dollar projects. Our pipeline is well over $10 billion right now. That said, we're not financing every single GPU that comes across our table. If we did that, guess what? Our pipeline would be north of $50 billion, $60 billion right now. The market needs to understand we are very picky, we're very choosy at the same time. Also, we have refined the way we look at project-level financing, whether it's structured debt, whether it's strategic capital, whether it's working capital facilities, infrastructure financing. We're bringing in the right people. I think you've seen we recently got the chief of staff.
We brought in a gentleman called Viral Patel in-house to help us with our project-level financing and structured debt. We acquired Shackleton, and now it's Gorilla Tech Capital. They are running around trying to find financing at the same time, what I think is long-term contracted infrastructure financing. Okay? They are going to be focused on this long duration. We're looking at five, seven, nine-year programs where we can attach the assets and the cash flow to the investments. That becomes particularly important for us because if you look at the program sizes, they will increase exponentially. Our objective is to make sure that we preserve the corporate liquidity, but at the same time, I also want to make sure that we're able to match what is happening in the market.
Now, for people who sit down behind their desks and start commenting about Gorilla, they need to understand we are building the opportunity, and the opportunity is about converting revenue cash flow into sustainable shareholder value. Personally, that's the only thing we're focused on in 2026, and I believe 2027 is really only about that for Gorilla.
Excellent. Well, Jay and Bruce, thank you so much for participating in this conference, and I want to thank everyone else for attending. More materials on Gorilla Technology are available on our website.