Gorilla Technology Group Inc. ($GRRR)

Welcome to today's fireside chat with Gorilla Technology. I'm John Roy. I cover technology here at Water Tower Research. And today, I'm joined by Jay Chandan, he's the CEO; and Bruce Bower, the CFO. I should mention that Gorilla's safe harbor statements can be found on their website. This fireside chat may not be reproduced or written transcript distributed without the expressed written consent of Water Tower Research. Jay and Bruce, welcome, guys. How are you doing?

John, absolutely fantastic. Thank you very much indeed. How are you?

I'm doing well actually.

So Jay, obviously, you made the Yotta announcement. Can you give us some of the color and insights as to what is in that announcement? And what does it mean?

Absolutely. The first takeaway, John, is quite simple. It's a moment where I believe, and Bruce, of course, we both believe that Gorilla has moved away from just being an AI company to an AI infrastructure company. Think about it from -- moving from PowerPoint to [ Power ], okay? For a long time, the AI market has been full of brand speeches. We've been looking at all of this glossy gags, you're looking at people. We're using the word transformational like confetti at a wedding. So for me, that's not what it is. Yotta is not about a pilot program. This is where real infrastructure, real GPUs, real deployment, but more importantly, real revenues. Now the initial Yotta deployment is about 640 for high-performance servers, and this is representing roughly around 5,000-plus GPUs. But if you look at this, what Gorilla is doing, it's moving from what we call the base case of reality to actually what really matters into what we can prove going forward. And that has already been what we call scaled up by the second contract we signed with Yotta, which we also made an announcement. That covers an additional roughly around 20,736 GPUs. These are B300 GPU cards, which will be the largest AI cluster within the region. And as you know, NVIDIA has taken 1/2 of the offtake agreement with Yotta. And all these have to be delivered by 30th of September this year. So, the total aggregate value of that second phase is roughly around $2.7 billion, $2.8 billion, net-net, putting about 225,000 GPUs with roughly around $3.2 billion, $3.3 billion of announced commercial values for the Yotta framework. Now for me, personally, that's not just a press release, that's intent. Now the second takeaway for us is India. Now India is not just a big market. India is one of the most important AI infrastructure markets on the planet after U.S., China, it's -- I mean, you're talking about 1.6 billion people with roughly around 1 billion-plus people using the Internet, roughly around 22.5 billion, 22.6 billion UPI transactions just last month alone. So you're not talking about a population story anymore. It's about a digitized population which loves data. And who loves data, AI does, right? Data needs compute. Compute needs power, they need cooling, they need racks, they need networking, they need security, and more importantly, people who can actually deliver the whole thing without turning this into a very expensive, what I call, glamorous bunch of radiators. So that's where we come in. We -- as you know, we actually have teams in Taiwan. We have a big data center team now in Thailand, India, U.K. and in Egypt. So we're bringing all of these together to create a global deployment team. The third one, which we are seeing is very important, which is the sovereign AI part of it. This is not just about compute -- the companies wanting compute. This is about nations where we are seeing countries like India decide that their data, all of their models, all of their citizen services like the Aadhaar Card and so on and so forth, their financial services and all of the critical infrastructure just cannot be sitting somewhere else because it is convenient, right? If India is building its own compute, as you know, they've gone from 38,000 high-performance GPUs, they'll be adding another 20,000 GPUs. We're also getting significant demand requests from India and the Indian government where we're seeing another 30,000, 40,000, 50,000 GPUs on top of what we've already built. So that tells you the direction of travel. It tells me that India is not just dabbling into it. They're investing billions so that they can be competitive. And finally, the most important thing is Yotta. Now Yotta is a phenomenal partner for us. It's not just like a local partner who just wakes up at morning and says, "Look, I've got a few racks in there and you've got a wave of flag, and thank you, we've declared victory." When somebody actually goes and sees the Yotta data centers, they will understand how serious these guys are. I'm talking about data centers that are 14 stories high. Personally, I have not seen anything of that scale today. So what they bring is they bring discipline, they bring the cooling, they need to bring all the customer access. More importantly, we've created a relationship of trust, finally. So all that, including our commercial execution, it's not just about performance, it's not just about a supplier relationship, it's about scale. So for me, the bigger message for your question, sorry, it's turning out to be a long response. It's about Gorilla becoming the picks and shovel company of the AI revolution, right? Soup to nut, the one throat to choke, as you said. Everybody wants to talk about AI, everybody wants GPUs, everybody talks about networks. For me, what Gorilla is building along with Yotta is, it does 3 things, which shows that the demand is real. It shows that India is enormous and Asia is growing as a banner. Third, Gorilla can execute at scale. We're not talking about smaller deployments of GPUs. So we are going to be banging our chests or beating our chest hard. We're going to make sure that those machines switch on. We're not just going to live on hope. But more importantly, I think this is where, for us, the fun begins. It's exciting.

Wow, that's a whole lot going on. So Bruce, how do you corral all this stuff to really focus on profitability and improve the balance sheet?

Well, the improvement in the balance sheet has been an ongoing focus for the company. So, at the end of 2024, we had about $21.5 million of debt. As of the end of the first quarter, it's down to $13.2 million. So it's been a very disciplined, sustained reduction. We didn't just go pay it off willy-nilly. It was basically when it came up for renewal or we didn't need it, we paid it down. And in so doing, we freed up restricted cash that was held against the debt. So basically, we had debt. We had to post collateral. The collateral was restricted deposit. So we pay off the debt, release the restricted cash. So net-net, it was very little unrestricted or free cash that was moving out to pay this debt down. And then we've also been disciplined in terms of reducing restricted cash balances. So just last week, we had an extra $5 million released. So, we're feeling very comfortable about the reduction in restricted deposits and debt. The cash that we have is about $90 million as of today on the balance sheet, and this is really with one goal. It's to fund all of the ambitious project pipeline that we have identified. So, I think the way we think about it is 2 ways. First is that there's plenty of cash for growth and then to take the steps that we need. So for instance, putting down deposits to place orders for equipment, et cetera, and also having a reserve so that there's plenty of runway for SG&A and everything else in the meantime. The other thing I would say is that profitability, well, all the projects that we're talking about, the new ones that we've signed up that we're implementing will be a boom to profitability. The sort of historical margin profile [ is between ] 30% to 45%. The exact kind of gross margin depends on the data center focused projects, but it's higher than the previous. So each of these new projects, when we implement them, improves the overall mix, and it's going to lead to improving margins, both gross margin level and then EBITDA level. So we're excited about the outlook for revenue growth. We're excited about the outlook for profitability and margins overall and also for cash generation going forward. And then the last thing I would say is that, as always, our focus remains on finding the funding where it doesn't stress the parent company balance sheet. So for many of the projects where we can, we're getting project level financing. So that means debt or, in some cases, non-recourse and non-debt financing at the project level. And then also where we're looking for additional funding, it would be -- we're looking at debt and nondilutive financing at the parent company level as well. And then I think Jay can talk about Gorilla Tech Capital and what that means as well.

Excellent. So we've talked about some of the recent future outlook. We've talked about controlling costs and getting the profitability better and that's stupendous. So Jay, give us an update of where you are on the progress? And how are you on the global rollout? I mean there's a lot of countries you've mentioned, but what's the progress?

Yes. John, I mean, you've covered us for a long time now, right? I think we've grown old together. So 3 years ago, Gorilla was largely seen as a strong technology company or a technology company with one market. Today, we're becoming a serious AI infrastructure player across not just Asia, Middle East, North Africa and even in Europe. That shift did not happen by accident. It was very deliberate. The decision we made was deliberate because we decided we're not going to simply sit in one corner of the world, polish a few products here and there and wait for the world to notice us. That, for me, was an expensive mediation with a bad return on investment, right? Now we decided that the demand is real. Governments are caring about sovereign AI. Enterprises need secured infrastructure and global players were not always moving very quickly. Some of them actually, we've realized move with all the urgency of the British passport office in August. So you can understand how close that is, right? But that has taken us to different markets, whether it's India, Singapore, Malaysia, Thailand, Indonesia, The Philippines, Taiwan. And with the Astrikos investment, we're actually now starting to look into the Middle East and United States. We're having multiple conversations, particularly around real-time infrastructure intelligence, smart infrastructure, and more importantly, AI-enabled operational control. That also leads us into that GPU deployment, service level operations, data center intelligence and so on. So we're not just dropping equipment in a room and hoping that someone is going to remember we have to switch the lights on or switch -- put the switch on. This is about real infrastructure where we're actually creating very complex machines. Now, this is also -- more importantly, it's about how we can help design, deploy and manage in these regions. Remember, when we are going into a region, we're not just building from, let's say, India or Thailand as a low-cost base. We're building within those regions. Like Indonesia, we're building a team now. Singapore, we're building a team. Malaysia, we just started looking at hiring as well locally. So when I talked about Thailand, Thailand is a great example. It's becoming a very important anchor for us. Why? Because we've been very active in Thailand. As you know, we're working in smart grids, we've worked in smart city, we've worked on data center security and critical infrastructure projects. With the announcement we made a day ago, Korat, that's where the story became real for us. This was about Gorilla owning land, Gorilla owning power, Gorilla owning fiber, water and execution at the same time. So the Korat, which is near -- in Nakhon Ratchasima is a campus that we recently purchased. It's roughly around 200 megawatts of total facility load, providing about 150 megawatts of IT load. There is -- obviously, we are working with the government to help us provide more power within that particular campus. And currently, the government is talking to us about additional 250 to 300 megawatts of power being available through the local substations, which are right next to the plant or the land we purchased. Now we're not talking about a proper AI infrastructure campus. We're talking about not a hyperscale but near hyperscale quality. okay? That means it gives us the ability to create high-density compute, it has GPU as a service, it allows creating sovereign cloud and hyperscale colocation. And really where it matters is it becomes a stepping stone into the next level as to how we're evolving. We're also looking at other land agreements in the region, not just here, but we're looking at Indonesia and Malaysia as well. And we're looking to invest into Malaysia at the Johor complex where we can actually buy land and build data centers. So across the region for me, it is absolutely clear. Singapore gives us the financial, commercial, regional discipline. Malaysia kind of gives us the, what I call, the power-rich data center market. Indonesia gives us the population. It gives us the scale and the digital demand. The Philippines is developing really quickly, and what is happening is this need for cloud, AI and sovereign infrastructure and the demand is increasing and accelerating quite rapidly. And Taiwan, of course, as you know, remains a core part of our DNA and which helps us develop all the products we want, whether it's the BMS products, whether it's your SOC/NOC and so on and so forth and more importantly, brings us the customer history. So in short, if AI for me was, let's say, a dinner party, so India would bring the appetite, Thailand would bring land and power, Singapore would bring discipline, Malaysia brings scale. Indonesia brings in population, but Philippines brings in the growth and Taiwan brings for me the engineering brain, and we have some very strong people. I think we have a little over 200-plus people in Taiwan. So, for me, now, when I go to any meetings, I look at it and say, great. Listen, wonderful speeches everyone. Now how are we actually going to build this bloody thing, right? And that's very important. So we're moving from country to country, but we're winning at the same time. It's not just about creating an office and a fancy place. So we are becoming aggressive. And as you know, a few years ago, everybody saw us as a technology company or a smart technology company. Today, people are seeing us as an AI infrastructure partner, which is helping them grow to the next phase, which is sovereign, secure and building regional GPUs. I hope that answers.

No, that's great. It does bring up the question, maybe one question, one final question is that, what are your plans for fundraising to support a lot of this stuff? And can you give us an outlook into 2026?

Sure. I think Bruce touched upon this very briefly. I mean this is a very touchy subject for a lot of people. Everybody thinks we're going to dilute the hell out of that. But let me start with this very important point. I think I need to let the market know. When people hear AI infrastructure, they immediately assume huge capital requirement, therefore, huge dilution, right? I follow some of the message boards, I stopped looking at them now because people keep thinking, God, this guy is going to raise $1 billion and dilute them. Frankly, some companies do that. And I can understand why they're upset by it. People get one contract, they run to the market, they issue equity, dilute the hell out of everybody else, very elegant. But it's like basically burning the furniture to heat your house. You don't do that, right? That's not our model. What we are trying to do is, our approach is very clear. We wanted to be very disciplined. And as Bruce mentioned, we're trying to be as nondilutive or with minimal dilutive structures. That is very important for us. Even if we're going to $1 billion, a couple of billion market cap, we will keep that rigor absolutely. That means we're looking at various options, project-level financing, customer-backed structures, asset-backed debt. Bruce talked about vendor financing, SPV-based financing, institutional co-investments. What am I missing? infrastructure funds. Then we have got strategic partners funds who are looking at AI data centers. We're looking at family offices who are looking at long duration of asset investment and not worried about the whole quarterly gimmick or anything as such. But the most important acquisition, which we did this year was the Gorilla Technology Capital, which is the Shackleton Finance acquisition with a full FCA approval. I think the market has missed the message on that. What people don't understand is that we have created a new route to fund data centers and GPU-as-a-Service deployment. How? We're basically looking to work with, like I said, SPVs, build co-investment, non-dilutive project funding across the infrastructure pipeline. And so we are targeting roughly around $2 billion to $3 billion of funds under management by the end of 2027. So the strategy for us is to raise money, but not to raise money for the sake of raising money. It's to make sure that we match the right capital to the right asset. So we've broken down our assets. So if you're looking at data centers, then we would be looking at infrastructure capital. People have asked me, how do you deploy GPUs? Isn't it going to be expensive? No, we're looking at asset-based or customer-backed financing. That's number 2. If you're looking at, let's say, long-term government contracts or enterprise demand like we have signed up recently with Yotta, then we're looking at contracted cash flow financing. That's number three. Then if you look at sovereign infrastructure, then we attract institutions using the likes of Shackleton, where we attract pension funds, endowments, institutions, strategic capital, right, and family offices. So for me, the lazy answer would be dilution. The grown-up answer for me is structure. And personally, I don't think so Bruce and I spent the last 3 years building or strengthening the balance sheet just so that we could hand over a dilution sandwich just because we have some financial jargons being thrown in there. Now, as you know, you asked about our outlook, we're looking at what, about $150 million to $200 million of revenue. Now again, that's another question for people. People are like, "Oh, my God, you've signed all these big deals. Why aren't you giving us guidance or re-upping your numbers, right?" I could do that. I could wake up one morning and just say I can do that overnight, but I have to be realistic. Data center projects do not happen overnight. The time taken to build the architecture, the time taken to work with the likes of NVIDIA, Supermicro, getting the architecture design correct, the networking and getting the supply chain. Nowadays, everybody is talking about memory and storage. People don't realize it takes 5 to 6 months to get it delivered. So how can I sign a contract and then come to the market and say, "Look, I have re-upped my guidance." I would be the absolute idiot unless the market is missing something. We're not -- we're building infrastructure. So, the most important thing is, 2026 for me was a year -- is a year, sorry, where we're going to be proving that the turnaround is real and where scale. It's no longer going from $100 million to $120 million or $130 million. We're going from $100 million to $150 million to $200 million. My personal target next year will be about $600-plus million, which I've already talked about. We said $500 million plus. So market, hear me out, hear Bruce out. We will adjust the guidance responsibly as and when needed. Data centers definitely take time to develop, deploy and finance as well at the same time. We're not talking about a microwave dinner where we can just cook it overnight. It's going to take -- we have to make the right level of guidance. We've got to get the market excited at the same time. It's not a 5-minute job. And I surely don't want to disappoint people at the end of 3 months and say, "Look, we've got to redo the guidance to a lower number." And then finally, again, my message to the market, we're not going to fund this like amateurs, okay? The last 5 months or now I've got 4 months and 5 days, I have spent more than 3-plus months on a plane, not because I fancy sitting on a plane and having my bones aching, it's making sure that we kept the funding. Bruce and I are on a plane almost every week. We want to make sure that we are not wandering around Wall Street with a begging bowl every time we need a cable. We're making sure that the money is real and that it is long term and sustained. So this is the difference between building a company or just a financing story. We're building a real company, John. So what is the outlook? Big opportunity, serious by discipline and pipeline and definitely no nonsense. So if you'd ask me, some people in this market are still doing this whole AI, like what is the best example I can give you? Like Mr. Bean trying to assemble an Ikea furniture without instructions, right? Lots of noise, one missing screw and a surely very worried set of investors. So we're taking the opposite approach, land, power, GPUs, customers, contracts, cash flow. Most important, cash flow.

Excellent. Excellent, gentlemen. Well, I think we're going to have to leave it there at this point. Jay and Bruce, thanks so much for spending time with me today and with the investors. Investors, to learn more about Gorilla, please visit our website or visit Gorilla's website. Our website is www.watertowerresearch.com. I want to thank everyone for joining us. And now I'm going to read a short disclaimer. The views expressed in this fireside chat may not necessarily reflect the views of Water Tower Research LLC and are provided for informational purposes only. This fireside chat may not be distributed or reproduced without the written consent of Water Tower Research and should not be considered research nor a recommendation. WTR is an investor engagement firm, not a licensed broker, broker-dealer, market maker, investment banker, underwriter or an investment adviser. Additional disclaimers can be found at www.watertowerresearch.com.