Good morning, and welcome to the Gorilla Technology Group's Third Quarter 2025 Financial Results Conference Call. [Operator Instructions] As a reminder, this conference call is being recorded. I would now like to turn the call over to our speakers today, Jay Chandan, Chairman and Chief Executive Officer; and Bruce Bower, Chief Financial Officer. Thank you. Please go ahead, gentlemen.

Thank you very much. Good morning, everyone. Q3 marks the strongest quarter in Gorilla's history with revenue ahead of expectations, operating profit firmly positive and the bottom line at breakeven. Now we've delivered a clear swing in profitability. We've built a cash position about over $119 million. We've reduced debt to a point of $15.1 million, and we've advanced our AI infrastructure programs across Southeast Asia, Latin America and the Middle East, securing multibillion-dollar projects, but at the same time, we're also creating a historic pipeline for this business. The simple message is that Gorilla is now operating above the analyst model and scaling faster than the market expected. Thank you. Bruce, anything you want to say?

Yes. I'd just like to take a walk through some of the highlights from the quarter and then in terms of where we are overall. So the first, as Jay mentioned, it was a record quarter for us in terms of revenue. The balance sheet, as Jay mentioned, $121.4 million of cash total. That breaks down to $109 million of unrestricted free cash and then the balance in restricted cash. Debt of $15.1 million means that we're in a significant net cash position of $106 million. This follows on the performance of the business and also in terms of the -- it was helped by a fundraise that we did in July. In terms of where we are as a business how and we're performing, you can see that we're on track to meet the guidance for 2025, which is in the range of $100 million to $110 million in terms of revenue. And then we were talking about EBITDA margins in the 20% plus range and net income margins in the 15% to 20% range. So we remain on track to hit all of those. The gross margins through the 9 months have been a bit over 35%. That's a little bit lower than we'd expect for the full year. So I think that we'll be on track to hit the 35% to 40% range for the full year. At the end of the quarter, we had accounts receivable of $36 million, and I know people are looking at that and worried. I'd just like to say that we expect the business to be collecting or some of those we've already collected on in the fourth quarter, a couple of significant outstandings in Asia and then some remaining in the Middle East, we expect to collect on. For the…

Thank you, Bruce. Yes, it was a very good quarter, rather, wasn't it? But if anyone is still wondering whether this is structural, I would gently suggest that they may need a new pair of spectacles. Now just to highlight on what Bruce talked about and clarifying some of the proof points to all the naysayers out there, our revenue, the consensus analyst model was roughly about $26 million to $26.2 million. Our actuals were at $26.5 million. Gross profit estimate was $9.5 million. We did about $9.9 million. Our operating income, IFRS operating income was to be at minus $6 million. We did a positive of $4.4 million. That's a big swing. And our adjusted EBITDA was about $5.6 million estimated, we did $6.8 million. Adjusted net income was about $3.5 million. We completed quarter 3 at $6 million. Our EPS non-IFRS was $0.26, and we came in at the Gorilla actual was about $0.257, which is in line. Our EPS IFRS was expected to be at negative 0.8. We completed it at breakeven, which is 0.00, which is a 100% improvement. Our analyst implied debt was at about $21 million. Gorilla's actual was at about $15.1 million, and we're looking to reduce that substantially before the end of this year. What we also had modeled for was the unrestricted cash, the restricted cash and the total cash position, and we are predominantly on top of everything today. Why? Because we delivered profitability at an operating level, not adjusted, not sprinkled with ferry dust, not if you squint, you can't see it, so on and so forth. This is proper profitability. We ran the business efficiently. We delivered on big projects across the region. We are delivering big projects across the region. We controlled our costs, but most importantly,…

[Operator Instructions] Our first question today comes from Mike Latimore from Northland Capital Markets.

Congrats on the great results here. In terms of the guidance for '26, what are you assuming on this large deal contribution kind of low end to high end of guidance? Or what are the factors that get you to the lower high end of that guidance?

Mike, good to hear from you. Let me answer it with numbers first, Mike. For 2026, we've guided a revenue range of roughly around $137 million to $200 million. This is built on only 2 things. One is our contracted backlog with very clear delivery milestones. Number two, the first phase of the Southeast Asia data center project, which alone contributes $100 million from '26 to '28. Now there is 0 revenue in that guidance from databases of the $1.4 billion program and 0 from any other new mandates that are being structured. Now the reality is that the remaining phases of the AI data center program are much larger than the Phase 1. As the time lines and the site consequences are finalized with the customers, we will then extend both our '26 and '27 revenue base quite materially as well. Now on top of that, as you know, we've also built a pipeline. Inside of these are several national projects in late stage that also touch data centers, public safety, network intelligence, our 5G offer inception programs and so on. None of that is in the current guidance of 2026. So the question you've asked me is the range we have given you is based on the backlog driven by a base case assumption. It is also dependent significantly on some of the very important issues we're facing today. One is material shortages of semiconductors, deliveries from likes of NVIDIA, Dell, HPE, Super Micro and so on and so forth. But that said, the upside from additional AI data center phases and new sovereign mandates will sit about all of these, and they will crystallize and therefore, they will become our future guidance as well. So I personally believe that we published a very sensible conservative number, and that's why we have deliberately left the rest out of them for now.

All right. Perfect. Any color on EBITDA margins, what you think they might do in '26?

Sure. Bruce, do you want to take that?

Sure. So we would guide for a sort of 15% to 25% range.

Okay. Good. And I guess just last one for me. The -- what -- can you provide a little more detail on the deliverables on this large contract in '26? Like what is the thing you're going to be delivering in the first quarter and throughout '26?

That's a good question. So the right way, Mike, to see the first $100 million is the run rate it builds. Personally, for me, I think most people expect that you've signed a $1.4 billion contract, it's a light switch and the revenue starts flowing in. No, it doesn't work that way. I'm sure you know data centers very well. We've been communicating on this for quite some time. The first one is basically about 6 to 8 megawatts. That's several hundred high-density AI rack. And we do not like them all in one day, as you can imagine. They come online in plan based as the power, cooling, all of the network zones are commissioned. So revenue ramps up in each batch as they are energized. Second, when you look at the GPU capacity, that becomes a very important factor as it follows in through these waves. As the racks go live, for example, we drop in the cluster through our NVIDIA and partner ecosystem, which drives up the GPU as a service usage line. Now on top of that, we stack our services over a period of time. So not all at once. You can't just do a big bang approach. It's video intelligence, say, for example, for cities, transports and borders, big data analytics, building your large language models for both the government and telco, bringing your inference engines and so on, your cybersecurity, your network appliances and intelligence platforms and things like even the environmental intelligence and smart policing. So as the national workloads move into the platform and the utilization grows, typically from 30%, 40% all the way up to, let's say, 70%, 80%, that deepens our revenue at the same time at the same levels as the physical capacity. So for -- just to take a leap from what Bruce said earlier, if you want us to be doing about $300 million to $400 million steady-state revenue, the GPUs all need to be in motion and be sinking harmoniously at the same time. So the part to that is a controlled ramp, as I said, is not a big bang. So we're anticipating, again, working very closely with NVIDIA on this. We're anticipating that we will get all of this commissioned and to go live by the end of 2026.

Our next question comes from David Williams from Benchmark.

Congratulations on the progress and success here, gentlemen. I guess maybe one of the first questions is kind of around the guidance. Obviously, you talked about this a bit earlier, but it feels like there is some potential upside there. And I guess if you kind of think about the risks in the market and maybe from the supply side and just the market dynamics, what do you think -- I mean, how would you gauge that from the midpoint of the guidance up to the upper end? And I would suspect that there's more upside opportunity than downside risk. Is that fair to assume?

David, it is absolutely fair to assume there is more upside. Why? Because see, let me give you the risks to the guidance. I think there are 2 parts to your question. First is the timing of the customer deployment. Large AI infrastructure and data center programs rely on client site readiness. There has to be site access, as you know, the market very well, power allocation, import clearances, customer procurement cycles, they can all shift from one quarter to the other. And even a slight change in a week or 2 changes that significantly. Number two, your supply chain constraints are also -- there is a big challenge today. If you look at the demand, there's a high demand for GPU servers, not just in the United States, but across the globe, right? And then if you're looking at things like networking equipment, they can also create longer lead times. I don't know if you've seen recently, the price of memory has shot up 40% in the last 2 months. Then you've got the things like regulatory and compliance approvals, you've got things like project phasing on multiyear platforms. You'll have to take -- we take into account even geopolitical sensitivities in certain regions like Southeast Asia, Middle East, Latin America and so on and so forth. But then if you look at the upside for us, I did talk about it previously. For us, it's about when these programs come live. Now our aim is to get all of these live by 2026 and make sure that we drop all of these clusters to our NVIDIA partnership and our partner ecosystem and make sure that we drive the GPU as a service usage line. Now once we've driven that -- and remember, these are all purpose-built data centers. That means there's one customer occupying 100% occupancy, okay? That means our revenue would hit scale as soon as the switch is switched on. So what we are trying to do is we are working very closely. I mean, I did mention to you the risks. But taking all those risks into mind, we're also looking at the upside. And we want to make sure that our upside actually helps negate the risks on the lower end. I hope that answers your question.

Can I add something to...

So a couple of other things, David, to keep in mind. The first is the data center opportunity -- the data center contract we have is an umbrella contract with Freyr. When we announced it, the $1.4 billion was based on the scheduled deployments that we had then. there is always the possibility that there are more deployments added to that. So that would be another source of potential upside. The second thing is, of course, while we're talking about the contractual backlog, and we talked about the data center side, we haven't talked about anything else. So Gorilla is still actively bidding for government contracts. And so we put in several bids recently, and we're staying tuned for good news from a couple of governments in Asia. The other thing is that we've talked many times about one Amazon and some of the MOUs that we've signed with government customers in the past. None of those are in the guidance now because they haven't yet turned into a date and an amount. But as soon as we know and have crystal clear vision on the date and the amount, then those would also be added to the guidance for next year. So it's not just about delivering everything from the data center contract, although that's the biggest mover. There are many ways for Gorilla to win next year.

And then maybe, Bruce, is there a way to size kind of the magnitude of your backlog? You've talked about a few things. You don't have the amounts or maybe even dates to. But if we were kind of thinking about your total backlog and kind of what you're anticipating for next year, how do you -- how should we size that?

So the backlog for us is -- we go with a strict definition. So $85 million is the backlog for 2026, where we have the exact date and time and it's signed and it's being implemented now. Then we have, as we mentioned, the data center contract where it's signed, it is being implemented, but the exact timing of the deployment is still -- we have a good idea, but it's not definite yet. As Jay mentioned, there are some [ DUCs ] that we have to get in a row or there are other people that we have to work with before we can define that. The pipeline is where we have a qualified lead, where we think that they'll make a decision in the next 3 to 6 months, where they have budgets, but that doesn't have a signed contract or with an amount and a date next to it. So there's 2 parts. One is the backlog is very strict. And then the pipeline for us is really about converting from customer either where it's signed, but it's not amount and dated or where they sign up and then they sign a contract and we know the amounts and the dates and can then move that into the backlog.

If I add some color to that, David, as well, the pipeline has grown rather enthusiastically, if I may. If it grows any faster, I think I might need to send a congratulatory card for myself. But that said, the deals are also very mature. If you look at what we did a couple of years ago and where we were last year, we were building POCs, we're signing MOUs and so on and so forth, whether it was part Asian in the U.K. or the Middle East, LatAm and so on and so forth. The data center project has accelerated beyond our expectations. And I don't want people to think that we're only building the data centers. There's a lot of ancillary support services we provide on top of that as well. So the $1.4 billion, for example, was only a catalyst. Once governments and telcos saw that we could deliver sovereign grade AI infrastructure, that basically kind of triggered a surge of interest. Now without giving names, the demand wave behind the FRR is significantly larger than Freyr itself. That is one of the primary reasons why our pipeline is well north of $7 billion. Now if you look at the GPU infrastructure, it has moved away from ambition for us to urgency. Through our engagements with likes of NVIDIA and Edgecore and including our own appliances within the kind of the government, we're seeing that strategic infrastructure as an essential, not optional. So now what has happened? We've also started working with the likes of Telstra in Brazil who's providing capital and looking to build some seriously large data centers as well. So these are all kind of whole country platforms as opposed to just incremental pilots. And then finally, what we are doing is that we're making sure that we can formally count a large portion, let's say, even if it's 20% to 30% of the $7 billion to be signed very quickly in 2026. And that allows us to actually be much more confident of our multiyear expansion. So in short, David, the opportunity is pretty comfortably substantial for us, but it's also growing at the same time. And it's not definitely a single year anomaly.

Okay. And one more, if I may here. Just if you kind of think about your competitors in the market and the 800-pound Gorilla, so to speak, you're competing against there. Why are they choosing Gorilla? What gives you the edge? And why are you winning?

That's a good question. Why we're winning? I think we've proven ourselves, okay, to where we are today. We believe that we work with governments to make sure that we understand what their requirements are, what their commission requirements are, what their ecosystem requirements are, and then we help them build national workloads. Now Gorilla has been in this space for a very long time. As you can imagine, we've been here for 24 years. We're going to be celebrating 25 next year, right? We are a full stack AI operator. And I think I kind of talked about this in my first speech at the NASDAQ, and I said we want to be an AI stack operator. We design the architecture, we build the data centers. We integrate the GPU stacks. We operate the platform, and we stay as a long-term partner for the governments and telco. Now apart from that, they also -- we offer these customers of ours, both enterprise as well as the government level, sovereign control and predictable economics. And that is very, very, very important because our customers know exactly who runs their infrastructure, who carries the responsibility for their uptime and performance. And then finally, it's about capability. Now as you know, we've been delivering national cybersecurity infrastructure. We built 4 data centers in Egypt for our $270 million contract. We're executing multimillion-dollar projects, national projects across Southeast Asia, Middle East, LatAm and so on. What has happened is we are moving faster than our competitors. Our speed of execution, our ability to structure these projects and our operational discipline is making us the preferred partner where you understand this probably better than most people do, AI infrastructure cannot fail. It does -- it cannot fail. It just cannot fail. And it has to be with people who can have a very strong operational discipline. I think that's the responsibility we take. So we will build, operate and manage responsibly. So think about it this way. Everybody is trying to sell buildings and servers. We're trying to sell outcomes. That's it. That's as simple as that.

And one thing to add on to that, as the numbers guy, is when I was investigating why we win, so to prepare some investor materials, all of that came out. The other thing is that given our history and our relationships with hardware vendors in Taiwan and then using our own software to create appliances out of the hardware, we actually deliver a significant cost savings over a competitor. I mean, obviously, the biggest cost item will be NVIDIA GPUs and there's not much flexibility. But on items where there's flexibility, we can deliver like a 30%, 40% cost savings with better performance, and that will reduce the overall cost of the data center by 5% to 7%. And 5% to 7% may not sound like much, but when you're talking $1 billion data center, that's a significant cash savings. So not only is it sort of everything that the customer is looking for in terms of sovereign data infrastructure, faster time to market, but it's also cheaper. So in the end, there's enough that stacks up, it becomes very difficult to look at a competitor by comparison.

Our next question comes from John Roy from Water Tower Research.

Obviously, a lot of discussion around '26. I want to step back for half a second and look beyond that. And kind of these questions are related. One is, do you need to grow your sales team to turn that pipeline into backlog? And can you give us some color on the pipeline beyond '26? And the last thing is, what are you going to plan to do with all that cash? Is it for growth? What's it for? Just kind of curious.

That's really, really good. No, no, that's a good question. You caught me off God there. No, but listen, I can tell you that my pipeline is $7 billion, and I can sign all of these deals, and it's all going to be hunky dory. It is not. It is going to take its own challenge. It's got its own challenges. Am I going to expand my sales team? Our sales teams are already well established. We have more than what, 250-plus people today. Full time, we have more than 200-plus contractors. So we're stretching our bottles right now. The sales guy -- there's one sales guy who gets everything done, which is myself. I make sure that I'm there in front of every single customer, every single project. It doesn't matter whether it's a $1 million project or a $1 billion project, I make sure that I'm there so that I can give them the confidence in the guidance. Where are we aiming for -- I think you kind of touched upon this as to what your -- what the future looks like. For me, personally, right, if the programs and partnerships in front of us land the way I expect it to be in the next, let's say, 3 to 6 months, I would like us to be -- and this is my personal target, please do not assume that this is going to be the company's target, around $500 million of annual revenue by '27. That's not a formal guidance, by the way. This is my target for what the platform is capable of delivering. Now that's what I am focused on. I want to get there, but we need to make sure that we've built all the LEGO blocks in place. to make sure that we're no longer a project shop, make sure that our pipeline is real and growing, make sure that we can have more cash and that it meets our ambition. And more importantly, it talks about what kind of acquisitions we're also able to do so that we are able to support. We need teams, we need people. Just to give you the scale, we've gone on a massive hiring free in Taiwan. Thailand is almost what, 60-plus people. We are looking at India. We've got about 150-plus new recruits going on in India. And we're looking at acquisitions as well for the first time in India as well as in the U.S. So that's -- keep your eyes peeled, and I'm sure we'll be able to provide you more updates in due course.

No, that sounds good. And the cash, maybe, Bruce, can you give us some highlights on where that cash might be headed?

Yes. So for all of the major contracts, there is a capital needs from Gorilla side. Sometimes with government customers, that can be for performance guarantees and for working capital. For some of these data center projects, we have to fund the CapEx upfront and then deliver it to the customer. In this case, we are in active negotiations with banks. I mean, Jay and myself are in New York this week, meeting with banks. So we have term sheets on the table from lenders, which will finance the vast majority of it. But just like getting a mortgage for a house, there's an equity component and the equity component would come from the balance sheet. We anticipate that we have more than enough cash on balance sheet now to fund the first deployment or 2 and hopefully even more than that. Like I mentioned, the business should generate substantial cash in the fourth quarter. And so that will see us into much higher revenue numbers in the coming 3 to 6 months.

Our next question comes from Brian Kinstlinger from Alliance Global Partners.

Congrats on all the business development achievements over the last few months. As it relates to as it relates to the Freyr contract, I'm curious or I assume the margins are substantially higher than the operating margin of your existing business. The offset is the CapEx side. So the cash returns maybe aren't what the EBITDA margins are, but the EBITDA margins are super high. I just want to see if my assumption is right.

You're right, Brian. First of all, good to hear from you. First of all, Freyr is not a construction gig. For me, it's a long-term AI infrastructure relationship across Indonesia, Malaysia, Thailand, Vietnam, Philippines and so on. Now what we are doing is we're designing, building, operating and monetizing it over the years. Now once that data center is live, we're not just there to host the racks. We're also layering a lot of services on top of it. So video intelligence, like I said, big data analytics for government, cybersecurity platform, smart policing and so on and so forth. So for me, Freyr is the doorway. The real value is what we sell on top of it and everything inside that footprint. So what we -- when you look at it from that perspective, yes, you're absolutely right. It carries a higher margin, your EBITDA is much higher. But in terms of cash generation, it might actually because of the CapEx -- extensive investment of the CapEx, it's going to be slightly full cycle. But what we will do is we will then deploy our own operations team. And more importantly, we will also apply our own stack of our solutions on top of them, helping them go from building large language models to inference engines and moving up the value chain going from let's say, H100 to 200 to GB 200, GB 300 and what comes after. And so look at it this way. For me, building data centers is only one part of it. Think of us as creating, curating, hosting and protecting your data. That's what we do.

Great. And then as we enter 2026, regarding your first large contract, which was the Egypt Smart City contract, how do you see the economics change in '26 versus '25 in terms of revenue? Are we increasing, declining, kind of steady state? And then how did the mix change from '26 compared to '25?

That's a really good question. So if you recollect about a couple of years ago, Brian, when we first spoke, I said my first job was to derisk the business. And it was to derisk our delivery profile in 3 ways. I mentioned this to you, and I'm going to stick to my guns here. First, we secured the contracted program. Once we did the technical validation with the government of Egypt, we then score our revenues and so on and so forth. And as you know, 95% of our revenues came from government customers. So what we did was we wanted to move away from projects to long-term milestone-based predictable collections so that our cash exposure is limited. It took us about 1.5 years to build that. And today, we're seeing that we're able to strengthen our balance sheet, but more importantly, we're able to reduce debt. Now what has happened, and this has allowed us to give us the breathing space to reengineer our business and to build our, what I call, capabilities at the same time. So look at it from having project-based schedules and programs to a full fledged deployment. These factors kind of helped us reduce our execution risk, revenue timing and financial risk. So going into '26, I can say with confidence that we are able to now have a more predictable, more stable quarter upon quarter as opposed to what we had previously. that answers your question?

Yes, somewhat. I'll take some of it offline. And then I'm curious, you had a number of MOUs, including Amazon One, there's a smart city contract. Any update on your progress? And I don't need to go over each one of them, but maybe where you're seeing more progress headed towards the finish line of any of the MOUs that are very large.

Yes. So the One Amazon project is going full steam ahead. As you know, we've already completed the proof of concept in Panama, and now we're running into Mato Grosso. You saw the signing happen sometime last month. So there's an initial $100 million program where we have received -- we expect to receive a good chunk of that in our tech deployment. Now of course, there are lots of issues we need to worry about because we have to worry about the sensors, the way they deployed, how every active is being monitored, how it becomes a stream of environment and health intelligence and so on and so forth. And these are monetized for decades. So we've already started work on that. It's growing, and that is not part of our guidance for 2026. We've also signed, as I said, with our MOUs with the likes of nTelastra, for example. This is not a single site. We're talking about 120-plus megawatts to be done over the next 24 months. So that also tie that to our Freyr project and so on and so forth, we're expecting that to also convert into a portfolio of other AI infrastructure projects, which are repeatable. We're also working very closely with the projects, and I know what is on the tip of the tongue of everybody in Thailand, for example, we are working very closely with the government and to give you some confidence that we are sure that there would be an outcome and light at the end of the tunnel over the next few months. In terms of the overall One Amazon and the other MOUs, which we've already signed, our team has been working day in and day out, and we're making sure that each of the platform builds the digital backbone and make sure that we are sitting on top of their infrastructure play. So again, all these are not included in the guidance for 2026.

Great. My last question, that was helpful. You highlighted, Jay, accurately that you invest to grow. You made a comment about that, and you've done that. But given the solid awards, the growing pipeline, are there any key investments you need to make now in terms of personnel, staff, facilities to take advantage of the opportunities in front of you? Anything meaningful that you can talk about or can quantify?

Absolutely. I think I touched upon this, Brian. This is very, very, very important because I think most people think that, no, we're a small company, we don't have the means to do what we do. So what we are doing right now, and just give it to you straight, right? These are all numbers back. So we are focused heavily on our M&A story as well because that brings in deep execution legs for us. But at the same time, we're also looking at how we expand ourselves into some of the fastest-growing economies in the world. So first, India. if you look at the India AI market today, and I mean, I may be slightly off on these numbers, but we're looking at about $9.5 billion today, and that's expected by 2032, I think, or '23, it's going to be about $130 billion. There's a tenfold expansion of the AI market. The country is also -- I was there recently, the country is also doubling its data center capacity from 950 megawatts to roughly around 1,800, 2,000 megawatts. By 2026, we're talking about a transformational change. So this is a massive national shift when it comes to AI compute cloud and digital sovereignty. So our investment into India is not cosmetic. Our acquisition potentially is also not very cosmetic. It positions us in a triple-digit billion dollar economy and where we are building our own local team, our own regulatory posture, but more importantly, we are looking at sovereign grade projects at scale today. So India is one big market for us going forward. The second market, which we talked about and which is also going to give us scale and people to help deploy in the local market is the United States. Now the U.S., as everybody knows, the largest AI market on the planet, represents roughly around 36% to 38% of the global AI spend today. But that said, it is also true that public safety, digital infrastructure, your GPU demand, defense and so on and so forth are all running into tens of billions of dollars apart from the data center market and the AI market. So for me, the acquisition there we're pursuing is very deliberate. It gives us established platform. It gives us huge customer potential. And more importantly, it gives us execution depth. And that's something you asked me to talk about as well to deliver, can I deliver real AI infrastructure and public safety programs in a country like the United States or India? This is how we're going to do it. So the U.S. for us becomes what we call a second engine for Gorilla for the next 2 to 3 years, not just a size project. So look at it this way. We're not buying revenue, we're buying capability. So that gives us scale. So India gives us scale in the hypergrowth market. U.S. gives us credibility in the world's most mature AI and law enforcement ecosystem.

Our last question comes from [ Bart Boone ] from Red Chip.

Jay, Bruce, congratulations on a great quarter.

Thank you.

I just have a few questions here. First, we know you design, build and operate AI data centers, provide GPU as a service and you're rolling out your own branded AI GPU platforms with partners like EdgeCore and Intel. At the same time, you're deepening your relationship with NVIDIA and the wider GPU ecosystem. So how should investors think about the unified flywheel you're building and Gorilla's strategic role inside the next wave of AI compute infrastructure?

That's a very interesting question. Well, I'll keep it short. The short answer to that is that we're not playing in one corner of the AI infrastructure. And I think the market needs to understand that. Why? Because we're building the whole engine. The data centers are just an anchor, [ Bart ]. We design them, we build them, we run them. We sit on them because they're long-term hosting and power and capacity contracts and so on and so forth. On top of that, we stack the GPU as a service using our NVIDIA-based platforms with our partners. That gives us usage-based recurring revenue as the workload scale. And this is a very important term, which the market needs to understand. As we scale, we will scale as well. And as our customers scale, our revenues will scale automatically. That term is called usage-based recurring revenue as the workload scale. Now on top of that, we talked about the flywheel. The flywheel is very simple. Data centers drive GPU demand. your GPU demand pull through our software, the software then locks in longer and deeper national engagement. Think of it as a 3-pronged approach. So how should someone see us, whether it's investors or customers, they should see us as a sovereign grade AI operator, not just as a project contributor or a box shifter. We're surely not a box shifter.

Thank you, Jerry. I think that adds a lot of color there. Now shifting away from the data center conversation. You've spoken about Quantum-safe networks and the Intelligent Network Director platform for lawful interception and network intelligence. How should we think about these as commercial gateways into larger sovereign infrastructure and national security programs rather than stand-alone products, right? How do they all work together?

The quantum question. I love that. [ Bart ], let me keep this tight. I know we're running short on time. This is one of the most misunderstood parts of our business. First of all, the market is enormous, right? Post-quantum cryptography alone is expected to cross over $100 billion to $150 billion globally over the next decade as governments upgrade everything from national networks to their financial systems to their defense communications and so on and so forth. Now look at this, every country will need this not want, but they will need it. That's an absolute must. Our Intelligent Network Director is never just a product. What we do is when a country lets you monitor its entire network flows, your lawful interception, your cyber posture, they're not just trialing a tool. They're effectively handing you the keys of their national nervous system. And this is what the market has misunderstood. We're not trying to sell a product. We're actually managing their national nervous system. Now that becomes a gateway into data centers, into sovereign cloud, into your public safety modernization, your AI workloads and your full national security stack and so on and so forth. Now as we move forward, right, the quantum-safe network opens the door even wider for us. Why? If you look at the way we protect country's backbone communications, we're automatically in the room. I mean, whether it's Taiwan, whether it's Thailand, whether it's Egypt, whether it's LatAm, it doesn't matter where it is. We are in that room for the next phases of their data centers, their GPU infrastructure, all of the national analytics, all of their secure workloads and all of their critical infrastructure protection. We signed 2 projects, as you know. And these were 5G lawful interception protecting national critical infrastructure. Now these technologies are the starting point for the programs that run into hundreds of millions of dollars over their lifetime. So what is Gorilla doing? We're sitting in that room. We're negotiating. We may sign tens of millions of dollars today, but my aim is to convert them to hundreds of millions of dollars over their lifetime. So think of it this way, whether it's your Intelligent Network Director or your Quantum-safe, we're not stand-alone. Think of them as a handshake that goes together over larger sovereign scale national infrastructure program. That's how I look at it from our IND perspective.

That's very helpful. I just have one more question to leave you with. So over the past few years, you've gone from survival mode to a position where you have record revenue, strong profitability, a multiyear AI data center mandate and a multibillion-dollar pipeline. What do you think the market is missing about Gorilla's trajectory when you look at the next 2, 3 years?

You put me on the spot there, about it. Okay. So first of all, I want everybody to understand this. We're no longer a project shop. we are becoming the sovereign AI infrastructure operator, right? Our Phase 1, for example, just in Southeast Asia, we're talking hundreds of billions of revenue per year. Later phases are just larger in scope. And the duration and none of that is in the guidance as yet. Again, I want to repeat that, it's not in the guidance. Second, our pipeline is growing. We're now sitting on our pipeline about what, $7 billion across telcos, law enforcement, infrastructure and government. These are multiyear national platforms. Now once we have proven that we can deliver, you're rarely a one contract supplier and the market knows that. Now if you look at the third part of it, balance sheet. And I think there's been quite a few questions on that. We have more than, like I said, $119-plus million of unrestricted cash -- sorry, $107 million of unrestricted cash and total of about $120 million of total cash left on the books. Now that means we can go fund serious data center builds without even blinking. A year ago, and you said it very rightfully, so we were managing survival. Today, we're designing national architectures. We're also making very clear, we're trying to make sure that we do not dilute our shareholders as a default. We're exploring a very range -- wide range of creative structures with our partners from project-level vehicles to revenue sharing and other funky options that let us scale hard without handing away the company to them. Now I did talk about my ambition. And again, this is my personal ambition, and this is not in guidance. This is not target. But…

We have no further questions. I'd like to turn the call back over to management for any closing remarks.

Thank you very much. Thank you, everybody, for taking your time and listening to us. To our institutional and retail investors, I'm going to say this out loud, and I haven't written this or practiced the speech before. Your conviction has carried us from survival to scale. Now people ask me about survival. This is very important. You stood with me, Bruce and the rest of the team through every single battle we have bought to get you. Now we enter a new phase. We're not just winning contracts. We're building the AI infrastructure of nation. Your belief has shaped this company, and it will definitely define everything we've been building in the years ahead. Most importantly, I want to thank every single one of you, naming people like Sam, people like Christian, people like Gunther, who actually stood by me while the world was still playing catch-up. And I intend to repay the trust with performance. So thank you. And thanks, everybody, for listening in. Have a lovely day.

This concludes today's conference call. Thank you for your participation. You may now disconnect.