Home / Transcripts / Gorilla Technology Group Inc. (GRRR) · March 2, 2026

Gorilla Technology Group Inc. (GRRR) Earnings Call Transcript

March 2, 2026

NASDAQ US Information Technology Software earnings 57 min

Earnings Call Speaker Segments

Jay Chandan executive
#1

Thanks a lot, Kristen. Thank you, everybody. I appreciate you being here. I'll keep this tight. If it's drama you're after, the market has served up plenty of that today already. Facts are what I'll focus on. Let me begin with the big news. Full year revenue hit a record $101.4 million, which is 35.7% higher than a year ago. Never before in our history had we gone past $100 million in annualized revenue; our guidance to the market was $100 million- $110 million. We landed within those bounds. That counts, since credibility counts. Our plan is to keep it that way. How we arrived here is the more significant piece; we pulled off a genuine turnaround. Our IFRS operating loss shrank to roughly $13.7 million, compared with $66.9 million a year earlier. That's a striking $53.2 million gain, meaning a 79.6% cut in the IFRS operating loss; meanwhile, our IFRS net loss came down to approximately $11.3 million from $64.8 million the prior year, an 82.6% improvement. IFRS basic EPS rose to around $0.51, up from -$6.13, marking a 91.7% gain. Indeed, that was quite a swing. It wasn't merely surface-level either; we accomplished all of it while preserving underlying profitability at scale. Adjusted EBITDA reached approximately $19.1 million, adjusted net income stood at roughly $19.9 million, and on that basis adjusted basic EPS was $0.89 with adjusted diluted EPS at $0.88. What I can say is that this is robust and highly disciplined. I already know the follow-up, since investors never fail to ask it, "How did we do versus expectations?" In the fourth quarter, market consensus sat at about $34.75 million in revenue with adjusted EPS of $0.30. Looking at our full year results, fourth quarter revenue came to approximately $35.6 million, comfortably ahead of consensus. Using the implied fourth quarter adjusted earnings, our adjusted EPS landed near $0.37, a beat of about 22% against that $0.30 consensus. Across the full year, market consensus was roughly $100.6 million in revenue with adjusted EPS of $0.84. We turned in approximately $101.4 million in revenue and roughly $0.89 in Adjusted EPS, a beat of about 6% versus consensus. My message is straightforward: Revenue set a record. We achieved a significant IFRS turnaround. And our underlying profitability came in above what was expected. Now let's discuss the wider market, since it's been turbulent. The conversation out there has moved away from whether you beat the quarter toward whether AI spending will endure. I'm certain you've all witnessed this over recent days and weeks. That debate is legitimate, though to me it overlooks the larger picture. AI isn't a discretionary software trend anymore; it's quickly turning into a national capability and a core operating layer for enterprises and governments. What comes next in AI demand can't be captured by 1 buyer or 1 deal. It will be captured by numerous buyers spanning different sectors, all building permanent capacity, governments, regulated enterprises, telecom operators, logistics networks, financial services platform. The list is extensive, and the spending has taken on a structural character. Compute itself keeps changing at a rapid pace. That's the piece the market is genuinely overlooking. AI compute is in fact moving away from a training-driven cycle toward one led by inference. That matters since it doesn't shrink demand; instead, it widens it. Inference brings AI into day-to-day workflows and mission-critical operations, which raises the requirement for distributed compute spanning regional data centers and edge environments, where latency, data residency, and resiliency requirements are significant. That's precisely where edge turns into a major driver. As the majority of you are aware, we ranked among the leading edge companies at the time of our IPO, and we keep investing heavily. Edge compute expands. What AI is capable of grows, because it brings inference nearer to the decision point, nearer to the sensor, nearer to customer interaction, nearer to regulated data. It turns into a force multiplier for adoption across public safety, transportation, logistics, financial services, telecom networks, industrial, and the entire plethora of smart cities. Let's discuss the scale of the infrastructure market in our region. We aren't sort of leaning on slogans; we're monitoring the data very, very closely. We have an internal team, we have a research team doing that work, and we draw on external data as well. We see Asia-Pacific data center investment rising from approximately $30 billion in mid-2026 up to roughly about $90 billion by 2030, 2031. We see installed capacity broadly doubling from roughly 29,000 MW today to approximately 63,000 MW by the end of the decade. Southeast Asia tracks that same path, expanding from the low teens of billions toward roughly $30 billion by 2030, as more capacity gets built in the market instead of being exported offshore. India is another illustration. It's scaling very rapidly. From a little over 1 GW of installed IT load today, they're heading toward about 1.8 GW by 2027 and to multiple GW by 2030. We're noticing the same trend across the Middle East. We're noticing a sovereign build-out dynamic, with market growth from low single-digit billions to a high single-digit billion by early 2030, as government and national champions scale local compute and secure infrastructure. What are we doing in 2026? We are pushing our AI infrastructure and data center build strategy forward well across Malaysia, Thailand, Indonesia, Singapore, and other regions, including Taiwan and so on. We're broadening our evaluation work in India. We're advancing our strategy in the Middle East, which includes Saudi Arabia, where an MOU has already been signed, and we're very actively exploring data center development opportunities in that region. We're also exploring opportunities to buy and/or build our own data center assets. Ownership changes the model. It hands us more control over our delivery and stronger long-term positioning and the potential to build recurring infrastructure-led revenue streams rather than relying on project cycles; in parallel, we are also strengthening our product edge for this next phase of adoption. The target date for having our SD-WAN with post-quantum cryptography on the market was April 2026, while our suites for lawful interception stay in ongoing R&D as we broaden sovereign-grade capabilities spanning security and intelligence plus deployments driven by compliance. Then in 2027, we are also assembling a team right now that will pour very substantial investment into lawful interception for 6G too. At present we have — what is it — roughly 300 full-time staff, and a bit above 200 contractors, all engaged on every project we have under contract. Looking only at the projects signed lately, we expect to scale up to somewhere near 1,200 to 1,500 full-time staff by the middle of June next year, which would mean adding somewhere in the neighborhood of 700 to 800 more contractors. That puts us at approximately 2,000 to 2,500 staff in the company at any moment in time. What investors are after is evidence; they want delivery, not storytelling. Let me address head-on the indicators that count: execution on delivery and on collections, and even more crucially, turning that into cash. Our biggest accounts are advancing quite powerfully, and how promptly we get paid shows how satisfied our customers are. Right? During January and February of 2026, we brought in over $22 million from our largest customers for solutions handed over and billed during 2025. We further anticipate substantial collections over the next few weeks.

Jay Chandan executive
#2

So we wrapped up 2025 holding roughly one-third of our assets in cash, totaling $104.8 million. What mattered greatly was that we accomplished this while bringing our overall debt down to approximately $13.8 million — that's a 35.6% decline compared to $21.4 million the year before. Then, by refinancing a number of lending arrangements and paying off others, we further cut our debt, which freed up over $5.3 million in deposits that had been sitting as collateral for some of those loan commitments. A balance sheet like this hands us substantial room to maneuver — to run our existing programs, to support working capital through delivery cycles, and to grow our infrastructure strategy carefully. At the same time, we've put more than $11 million toward buybacks so far, because in our view the market keeps pricing Gorilla below what our results and our plan deserve. Honestly, you might describe that as confidence. I'd describe it as math, wouldn't you? And that brings me to what's next. Our target is to reach positive cash flow in 2026; for me that isn't merely a tagline; it's a operational goal requiring rigorous execution, tight cost management, and strict discipline around collecting what we're owed. Last, many of you have raised this same question repeatedly. Gorilla Technology Capital. From where I sit, it's a transformative accelerant for the coming chapter. Its purpose is to broaden what we can take on in bigger infrastructure projects by arranging capital smartly, matching long-horizon funding with long-horizon assets, and helping our clients move more quickly through transparent financing routes. Some folks wondered, "Perhaps they're acquiring the bank." No — acquiring the bank is not what we're doing. You all need to grasp the function of Gorilla Technology Capital; it boosts our capacity to scale data center construction, speed up GPU infrastructure rollouts, and, even more crucially, take meaningful roles in bigger mandates featuring institutional-caliber structures and governance. To capture 2025 in a single sentence: we hit a record revenue landmark. We pulled off a significant profitability reversal; we fortified the balance sheet and readied Gorilla for the coming era of AI infrastructure — one that is sovereign and regional, and even more crucially, distributed, increasingly moving toward the edge. Come 2026, we move from demonstrating that we can deliver to scaling up what we deliver, turning execution into cash, broadening our data center presence throughout India, Malaysia, Thailand, Singapore, Indonesia, the Middle East, and, even more crucially, leveraging Gorilla Technology to open far bigger opportunities without any compromise. All of this while speeding up our product roadmap, meaning we're putting substantial investment into R&D. I appreciate your time. I'll pass things to Bruce, who is familiar enough with the figures to rattle them off without flinching. Bruce, the floor is yours.

Bruce Bower executive
#3

Thanks, Jay. On the financial side, I believe you hit the key items. There are a few things I'd like to touch on. To start, we noted a cash balance of $104.8 million at year-end. I want to stress that thanks to collections received so far this year, that cash balance has actually gone up. On February 26, unrestricted cash stood at $108 million, with total cash of $116.6 million — and that's even after we spent $3 million during this calendar year, meaning over the past two months, on share repurchases. So we've managed both to grow cash and to buy back stock this year. That's a solid beginning to the year. Another item I'd highlight: when we discussed loosening the debt burden, or lowering it and unlocking cash deposits, some folks asked, "Why not retire all of the debt?" Well, what remains — the $13.8 million — carries an average interest rate of 3%. Frankly, it's smarter to hold that as flexible capital rather than pay it down and then borrow at higher rates. The final point I'd mention: last year we issued revenue guidance for this year in the range of $137 million-$200 million. We're keeping that in place. Right now, we're not ready to provide gross margin or EBITDA guidance, but stay tuned over the coming months. We explained the reason for such a wide band between $137 million-$200 million: it hinges on the delivery schedule for certain data center projects we're pursuing with FREYR, as well as with others. I think we'll have a very good update in the next month to a month and a half regarding the timing of those projects, the delivery schedule from NVIDIA, and also with the customers; that ought to help solidify the guidance and give you a clearer picture. With that, I'd like to echo what we said in the press release and what Jay said. We feel the balance sheet has strengthened enough that we can fund our growth initiatives and also repurchase shares when we believe they're undervalued. We can take on many of the growth projects we've discussed — not only the revenue increase this year, which at the midpoint of the range would be, you know, nearly a 70% increase, but also the contracts in our pipeline. With a $7 billion revenue opportunity in that pipeline, we believe we can fund a substantial portion through access to debt facilities, primarily via project finance, and then through the cash currently on our balance sheet. With that, I'd like to hand things back to Jay and, you know, if you want to open up for questions, we can do that.

Jay Chandan executive
#4

Appreciate that, Bruce. I want to invite anyone who's waiting on the line to jump in with their questions now. Thanks so much.

Brian Kitzmantel analyst
#5

Yeah, that's about right. Excellent. Thank you very much, everyone. Over the past 2 years, you've certainly traveled quite a distance. Congrats on that.

Jay Chandan executive
#6

Great.

Brian Kitzmantel analyst
#7

Regarding your best estimate for when the first three phases of the FREYR partnership will happen, has anything shifted? My understanding was that project financing would let you kick off phase one in April, phase two in September, and phase three in December. Also, apart from funding these projects, do any other gating factors stand in the way of getting them started? And if there are, what has to occur during those windows?

Jay Chandan executive
#8

Brian, great to hear from you, and thanks for the nice words; where we stand today, we remain on track. Clearly, given current market forces, delivery has slipped somewhat. That said, let me sort of take you through what has taken place. Certain programs have shifted in their timing, and we discussed the FREYR contract, as an example; that one remains on schedule. Right now we are in the last stages of receiving our initial batch of GPUs within the coming days, and we'll be deploying it as we speak. We've also sped up the timelines on a number of data center conversations. During our previous call, if you recall, I believe we were considering roughly 12.5 MW of data centers. With that, you know, we were contemplating approximately several high hundred, high density AI racks. What we chose to do was, instead of bringing them all online on one day, we're gradually placing them into plan base. You know, power cooling network zones have all been commissioned. Revenue ramps will be energized as we speak. As racks go live, we will insert the clusters via our partner ecosystem, which additionally fuels the GPU-as-a-service usage line, as I call it. What's really thrilling for us, and I can share with you today, is that we've now come to see we would need to kind of put a great deal of capital into the data center space on our own, since we're being flooded with an enormous number of requirements. At present we're looking at roughly more than 600 megawatts of capacity instead of the, you know, 12 megawatts alone. That lets us control our destiny across a period of time, meaning we're looking at several hundred million dollars annually once all these racks and the GPUs are all in motion. From where we sit, Brian, the route to that particular point is a highly controlled ramp up, not one single bang. You asked about, you know, whether any delays exist? So far there are no significant delays. You know, Thailand MoE, as an example, was delayed owing to the political transition, some kind of departmental organization. You know, with the new Prime Minister now elected, we're simply waiting for the post-election leadership and sign-offs to settle as we speak. You know, apart from that, we aren't facing any delays. All the approvals, permitting, site readiness, and customer prerequisites are being pushed forward, lit as we proceed into it. Once those customer gates reopen now, I believe billing will kick off on schedule. Hopefully that addresses what you asked, Brian.

Brian Kitzmantel analyst
#9

That answers it, thanks. For my follow-up, you mentioned having a big pipeline of additional data center opportunities.

Jay Chandan executive
#10

That's right.

Brian Kitzmantel analyst
#11

I wouldn't call your business development sluggish — it's actually been extremely rapid. Do you figure those clients are holding off until they can observe how the first FREYR contract gets executed? Will that, in the short run, prevent agreements from being signed, or do you believe they could proceed regardless of whether those three projects get delivered?

Jay Chandan executive
#12

Definitely not. As I said earlier, the opportunities coming our way are multiplying rapidly. Our sales plan hasn't been sluggish at all, I assure you. The only thing we've done is exercise restraint. We've been flooded, and that's no overstatement. Flooded is precisely the correct term. To begin with, the deals have reached maturity. Back at the beginning of this year, in January, we were examining POCs and MOUs and similar arrangements. Those looked highly promising, and since that time we've progressed into late-stage commercial structuring or into fully executed contracting, which by its nature tends to boost both the magnitude and the reliability of what's in the pipeline. If you remember what I said at the close of December, around late December, we're ensuring that the pipeline has certainty behind it. As I noted, that $1.4 billion Southeast Asia contract served merely as a trigger. Once telco governments essentially witnessed what we can deliver, and we began signing on, you know, that initial 12.5 megawatts, all at once, out of the blue, it set off a kind of sovereign grade AI infrastructure demand and an enormous spike in interest toward us. I'd rather not name names, but what's occurred is that the demand underpinning this is considerably bigger than FREYR on its own. That's among the main reasons our pipeline now sits in the billions of dollars. The third and most crucial aspect is that things have shifted away from, you know, mere ambition. Governments no longer view it as an ambition. For us, it has become a matter of urgency. I brought this up last time too. Not only do we regard, you know, GPU capacity as strategic infrastructure, the transition has in fact advanced into edge compute. Distributed environments are currently taking form. As I said, the market already overlooked that. People wonder, "Oh, will the spending keep going?" It's going to speed up. It won't keep going at its current pace. It's going to become exponential. We're engaged with every single major customer on the planet. I can tell you, these platforms are simply going to explode when it comes to compute requirements and demand. Lastly, our execution hasn't been limited to just one data center, Brian. We've been building data centers for a very long time. We constructed data centers on behalf of governments, for instance, in Taiwan, in Thailand, in Egypt, and so on and so forth. Take situations such as rolling out lawful interception efforts at a large scale — those are harder to pull off than simply standing up a data center. Governments and organizations examine what Gorilla has produced, and they say, "See what Gorilla got done." Once they witness that, trust builds. We aren't sitting back on what we achieved before; instead, we're getting everything moving. As I mentioned when I answered earlier, our aim now is more than 600 megawatts of power. The chance in front of us is quite comfortably sizable, Brian, and it keeps expanding.

Brian Kitzmantel analyst
#13

Excellent. For my final query — you mentioned the people you need to bring on board. I appreciate that; throughout my professional life, in a company like yours, that tends to be a superb early signal. How would you describe the talent acquisition landscape across the locations where you're adding staff? Beyond those doing the execution work, is there a substantial requirement at the senior executive tier for AI HPC, giving you that added strategy and know-how at the top level?

Jay Chandan executive
#14

Ah, that is an excellent query. As you're aware, we're bringing people in very quickly. What isn't visible to you is, well, on our site the identities of the senior individuals we've already brought aboard. Over in Thailand, we're truly gaining momentum with, you know, roughly 80-plus individuals added. Over in Taiwan, we've established a considerable data center group plus an R&D group dedicated to our cybersecurity offerings; we've accomplished this using something known as a hub-and-spoke model. That matters a great deal since our R&D platform alongside engineering must speed up both our product and our services abilities. Regarding services, as you're aware, Satish joined us midway through last year, and he's been pushing all of the client impact and strengthening our technical abilities. Regarding R&D, we've been bringing on people for SD-WAN, post-quantum cryptography, lawful interception capability, video analytics. We've been expanding the product. As I pledged, by April we would have completely launched the world's first fully ready post-quantum crypto SD-WAN. We're already engaged in enormous proof of concepts with customers too. This is the thing that counts, Brian: localization. Each region we operate in, whether that's India, Middle East, North Africa, Southeast Asia, East Asia, all of them are inquiring about how you're establishing stronger ground capacity. What's our approach? We're assembling teams in Thailand. My group in Thailand, for instance, since we're examining some very large data centers here, will number about 1,000 individuals by the close of this year, it'll likely be 1,000 individuals in Thailand; we'll have somewhere between 200 to 300 individuals in India. Our Taiwan team will exceed 200-plus individuals. Now, we're bringing on senior executives too at the same time. As you've observed, Thomas has arrived and joined as a CTO of infrastructure. Jackie has arrived from the hardware side and become the GM for Asia. We're also bringing on next-level capability beneath them too. Simultaneously, we're also ensuring that finance and compliance get tightened as well. We're bringing people on to enhance cash discipline, collections, control, audits, and so forth. Consider it this way. The hub-and-spoke model is going to be centered across each of these regions, and as we expand and grow, we will be scaling our teams quickly over the coming few months. The teams are all set and operating at a fast pace.

Brian Kitzmantel analyst
#15

Great. Thanks for all your answers.

Jay Chandan executive
#16

Thanks, Brian.

Brett Knoblauch analyst
#17

Hello, and thanks for the presentation. I have a handful of questions. First, regarding gross margin, I'm curious about the revenue mix, since it produced a gross margin somewhat different from my expectations—I'd just like to grasp what that mix of revenues consists of. My second question: since you're going to roll out the newest compute for data centers across Southeast Asia, what revenue level per megawatt are you projecting there? And what kinds of use cases do you have in mind for it?

Jay Chandan executive
#18

Bruce, would you like to handle the first part? I'll take the second.

Bruce Bower executive
#19

Absolutely. Absolutely. I'd say a more useful frame is this: during 2024, the mix of revenue skewed unusually heavily toward service — service made up the bulk of it — and then in 2025 hardware accounted for a bigger share. It landed at roughly 40%. That explains why gross margins came in somewhat beneath what you'd anticipate. Beyond that, we shared last year that we'd brought on two significant law enforcement clients in Asia, and for at least one of them, the margin we'd projected when we took on the project was somewhat under what we'd typically agree to; the reason is that it represented an important win for us as a client and, you know, as a solution that let us show what we can do. Taken together, that's the reason the margin slipped down somewhat. Looking ahead, and picking up on what Jay said regarding the pipeline, I'd put it this way: we're in a position to be extremely selective about which projects we take on. Given how much demand we have, you know, if the margin terms, if the credit terms or the customer's credit profile aren't suitable or, you know, if the payment terms aren't in place, then we simply respond, "I'm sorry, either you get in line or we'll go on to the next project." Another point: the data center, the GPU-as-a-Service, carries a very high gross margin, so it's 70%+. 70% is essentially the floor. Clearly there's a depreciation impact, since an SPV would hold the equipment, which then gets consolidated onto our financial statements, and we'd record the depreciation charge. At scale, I'd describe that as roughly a 25% operating margin. That's at scale. I'm not giving the margin forecast for this year just yet; we're going to hold off until the exact details are firmed up; that's how I'd view, you know, 2025 as something of a dip when it comes to gross margins, and I'd expect them to improve over time and, you know, a considerably stronger margin profile across all the new business we're bringing in.

Jay Chandan executive
#20

Brett, let me jump in on that too — and what matters even more is that we are putting an enormous amount of capital to work, you know, to build this business for durable growth over the long term along with gross margins, correct? That leads me into the other half of what you asked. When it comes to how we price things right now, you know, over in Asia, the structure is either what we refer to as capacity per server per month, or it's based on usage per kilowatt hour, and which one applies depends on the customer and the program. Typically for sovereign enterprise rollouts, our preference is a multi-year contractual take-or-pay structure, with pricing, durable margins, and cash conversion all locked in ahead of time, so there's no ambiguity about what we're committing to. Quoting one flat number isn't something we do. Personally, I won't put out a single rate, since as you're aware it shifts depending on GPU class. Contract duration, how the capacity gets used, power and cooling requirements, geography, the worth of the land, the service level stack, and many other factors all play in. We only have real validation once we've inked these agreements, where the unit economics stay tightly controlled and our cash position is shielded by collections and milestone payments. If I could put it this way, there isn't one uniform rate for Asia. And by the way, our focus isn't only on Southeast Asia. No single Middle East or Asia price exists either. That being said, what I can share is that normally, for CSP class GPU rack capacity, the range tends to be high four figures up to low five figures for each GPU monthly, once power, floor space, connectivity, and managed services are bundled in; keep in mind as well that these agreements are long-dated and fixed to milestones. Quite often we add layers, which we refer to as service level fees, compliance components, and so forth. Any one of these is subject to change. Take the U.S., for instance — you'll generally find that on-demand pricing for high-end GPUs runs somewhere between two and three times what structured regional capacity costs over in Asia. Our approach doesn't involve setting one flat rate. Since the compute specs we need are tighter and the agreements we sign stretch out far longer, we can land on pricing that I'd describe as extremely competitive — even when stacked against the United States. Here's one way to look at it: a premium for compliance plus a premium for service accounts for roughly 20% to 40%, and baked into that are things like governance, telemetry, managed operations, and similar items. Because of the way energy prices differ, the arrangements we make in Asia tend to bring in considerably better returns. To put it another way, putting Asia and the US side by side is a bit like assuming a Vegas hotel room will cost less, when in reality a Bangkok penthouse can run you more than certain ones found even in Manhattan. That's my response to what you asked, Brett.

Brett Knoblauch analyst
#21

Yes, definitely. I appreciate it. Could I quickly jump in there?

Jay Chandan executive
#22

Go ahead.

Brett Knoblauch analyst
#23

Two last items, both minor. About the Astrikos deal you closed — from a strategy standpoint, do you plan to establish, or have you already established, particular pricing and margin targets for the fresh contracts signed in connection with it? At present, is it being packaged together to strengthen your competitive position and improve metrics such as long-term customer lifetime value?

Jay Chandan executive
#24

Excellent question. Allow me to share a bit of an update regarding the rationale for our investment and the reasons we're integrating, correct? My assumption is that's what you're getting at, together with our intentions. How would your springboard appear, right? Were you to ask me, I'm aware that's something I'd pose to myself. So, when we took a closer look at Astrikos — and let me start with what it actually is. Astrikos functions as an intelligence engine for infrastructure in real time, handling monitoring along with prediction and optimization for systems that are critical. It's already out there working in quite demanding settings, and that includes smart city platforms operating at high state levels. Take, for instance, the new Indian Parliament Complex — Vijay and I have discussed this before. There are also significant efforts underway in the Middle East. Why does this matter? Because Astrikos isn't some demonstration — it's a solution fully deployed. Now, on the second half of your question regarding our plans for it: we are folding Astrikos into three areas of our stack. The first, and the most significant one, is smart city and national infrastructure operation. This provides us with both a telemetry component and a predictive capability, allowing national infrastructure to become something we can measure far better while also optimizing it live. How should we understand that? It boosts our capacity to offer outcomes rather than merely hardware or software — things like genuine uptime, response times, threat identification, and ultimately what people refer to as delivering strong operational efficiency for the client. Next comes video intelligence along with security. Astrikos, as you're aware, generally makes announcements regarding real-time monitoring, or positions itself around critical infrastructure, security, and operational workflows. It pairs well with our video intelligence offering, and it enables us to better operationalize data throughout our SOCs and NAC environments. Lastly — and this matters a great deal — this one is significant. When it comes to data centers built for GPUs and similar environments, a typical facility simply isn't capable of handling really demanding GPU workloads. What you need is ongoing telemetry, optimization driven by prediction, security that's integrated, and automation of operations. That's precisely the need Astrikos addresses. Looking at it as a launching point — from my perspective, Astrikos serves as a springboard in India, and the impact is felt right away, since it gives us a strong regional presence, cuts down how long our sales cycle takes, and makes us more ready to deliver. Over in the UAE, we're already at work establishing our footprint across the Middle East. When you look at the U.S., it operates as a market where partnerships are the norm, so we're moving ahead with work at the market-entry stage in that area too. Think of it this way: we hold the role of a meaningful minority investor. We have an option on the table that would allow us to boost our ownership by a large margin, and it would still leave us with ample room to maneuver when it comes to folding in and pushing forward this transaction at a truly major commercial size. Brett Knoblauch?

Brett Knoblauch analyst
#25

Yes, definitely. Thanks. That really is quite useful. I appreciate it a lot.

Jay Chandan executive
#26

Thanks, Brett.

Michael Latimore analyst
#27

Okay, yeah, I really appreciate it. Yeah, congratulations on such a strong year. The results were excellent.

Jay Chandan executive
#28

Thanks.

Michael Latimore analyst
#29

I suppose only a handful of items. You mentioned that perhaps additional collections might arrive during this quarter. Could you give us a bit more context on that? Are we discussing a few million dollars? Are you referring to more than $10 million or, you know, or perhaps you're unable to comment, but just somewhat interested in that.

Jay Chandan executive
#30

Bruce, would you like to handle that?

Bruce Bower executive
#31

My estimate would be roughly 10 million dollars, give or take a couple million — so 2 million to 3 million in either direction.

Michael Latimore analyst
#32

All right. Is that tied to what you're doing in 2025?

Bruce Bower executive
#33

Yes — these are solutions that got delivered and invoiced during 2025.

Michael Latimore analyst
#34

All right, terrific. I just want to keep this straightforward on my end. Regarding that sizable Southeast Asia deal, it sounds like there's essentially no difference when it comes to the overall amount or the amount for each of the initial three data centers. Is that accurate?

Jay Chandan executive
#35

Yes, that's right. As I said earlier, that has turned into a catalyst.

Michael Latimore analyst
#36

All right, excellent. Jay, you mentioned possibly receiving your initial batch of GPUs within the coming days. I suppose I'd like a bit more detail on that. Is this tied specifically to the Southeast Asia deal? And also, were you essentially saying that you anticipate getting some of these GPUs on a weekly basis, and then that accumulates gradually? Perhaps a bit more clarity on that kind of pattern.

Jay Chandan executive
#37

Certainly. I believe we're building a flywheel effect, if I may, Mike. What we're doing is ensuring that deliveries keep coming in each week. Our most recent agreements with our OEMs provide that beginning next week, we'll have a handful of deliveries arriving. Again, well, I've said this before too, we've actually secured additional contracts as well, so we're actually fulfilling those contracts too. You'll observe a steady stream of, you know. That's the reason we've brought on a very strong procurement team too, which will ensure these deliveries arrive on schedule. For us, you know, these data centers are fueling GPU demand, and for us, our GPU demand opens up much more substantial national engagements. You shouldn't view the FREYR contract as a one-time thing; this is actually, as I said, a stepping stone to some very large contracts we've already signed; we've also agreed, incidentally, with all of the OEMs, local OEMs in the region. We have signed every MOU that's needed. We've signed all the LOIs and the pricing agreements. The BOMs have been finished. The SOWs have been completed. As you know, we're now just working on the delivery schedules and the mechanisms over the coming weeks.

Michael Latimore analyst
#38

Understood. So it sounds like these GPUs are headed to customers beyond the one in Southeast Asia?

Jay Chandan executive
#39

That's right.

Michael Latimore analyst
#40

All right.

Jay Chandan executive
#41

Could you grant us just a handful of extra days on that one? I'll also provide you with a highly specific timeline.

Michael Latimore analyst
#42

All right, that's helpful. So regarding the Southeast Asia deal and that initial data center, is your expectation still that it becomes operational during Q2?

Jay Chandan executive
#43

Yes, that's right. We're actually attempting to accelerate it into Q1, depending on how delivery schedules line up. But I have complete confidence — 100%, actually 101% — that it will be up and running in Q2. The BOM agreement was just finalized. Those BOMs have now gone out to our OEM partners. As you'd expect, it isn't only GPUs that need to arrive; a substantial amount of networking gear has to be delivered alongside them. As we scale alongside the customer and demand picks up speed, we'll need to keep layering more on top of that. Mike, one point your question brings up is something else that matters a great deal. We've been having difficulty meeting all of the compute demand coming from our end customers across these regions. As you'd expect, the US is putting hundreds of billions of dollars to work. That level of investment isn't something we're seeing in this region. Sure, KKR recently bought STT for $10 billion, but even so, deploying data centers at scale requires far more compute. Internally, we made the call to construct our own facilities using modular technology. Right now we're aiming for roughly 600+ megawatts, and with any luck, we ought to finish signing all of that before this year wraps up too. We'll also be moving into full-scale production in the later portion of this year. We're incredibly excited, and we believe we're effectively opening up a brand-new market that doesn't exist at present.

Michael Latimore analyst
#44

Great. Perhaps the plan to acquire and develop a few of your data centers alters this question. On your business update call back in January, I believe you said that you're attempting to lease whatever capacity is available in co-locations throughout the regions. I suppose, is there any update on additional new leases you've completed?

Jay Chandan executive
#45

Yes. Yes. Yes. We've already signed numerous deals in the region. It's truly fascinating. The issue is, as I mentioned, a lease that isn't there. Whether it's 9 megawatts, 4.5 megawatts, 9.9 megawatts, 21 and 25, that's roughly the capacity available at present. Okay? So, you're completely right. What will we do? We're just going to pivot and construct fresh capacity and deliver infrastructure ourselves to the end customer, correct? Mike, I haven't clarified this before. Our demand sits in hundreds of megawatts, okay? Asia, not merely, I'm not referring to Southeast Asia or East Asia or even South Asia. Asia-Pac overall lacks the capacity at this moment. India, for instance, has just 1 gigawatt of fully utilized scale. As you've seen lately, they held the AI Summit, and India is going absolutely bonkers when it comes to deploying the scale. There are other structural issues; we need power, we need water, and so forth, right? We are working, and just FYI, we are working very closely with the Indian government to ensure that we get our infrastructure ready across various requirements and various architectures and edge deployments in the country as well. Long story short, keep your eyes on your field; we are definitely headed in the right direction over the next few days.

Michael Latimore analyst
#46

OK. That's great. Thank you so much.

Bruce Bower executive
#47

There's one point I'd like to tack onto that.

Jay Chandan executive
#48

Thanks, Mike. Go ahead.

Bruce Bower executive
#49

When we're considering whether to set aside capacity, get it lined up, or construct it on our own, that's an entirely different situation. Building at scale and just hoping buyers show up isn't what we do. Our focus is on purpose-built, AI-focused data centers, or GPU-as-a-service, for those clients. So what does that mean? Number one, no capital gets committed by us until there's obvious demand from customers. Number two, we require prepayments from customers, because, as you know, money talks. After that, in the majority of situations, those customer prepayments form a key piece of how we finance things, so that combining project finance with customer prepayments, we're able to cover 90% or more of a project's CapEx cost. What we've discovered is that, you know, when customers demonstrate commitment upfront, it clearly gives us greater comfort in proceeding, and it additionally raises the odds, you know, that the economics turn out favorably for us.

John Roy analyst
#50

Thanks. Clearly, a few developments took place over the weekend. Could you provide us with some sort of update regarding operations, or perhaps the outlook for the Middle East, in light of the situation between Iran and the US?

Jay Chandan executive
#51

Mr. Roy, I appreciate you raising that. To start, for everyone listening and everyone out there, I truly feel awful about what's going on. My heart goes out to every family impacted by this, and to all those who have lost people they love. I feel deeply, deeply saddened. I have friends on both sides of the Atlantic; and from a business standpoint, John, we're tracking developments extremely closely. As you're aware, our approach to risk is extremely, extremely, extremely disciplined. Right now, no meaningful effect is being seen on any of our operations. Egypt is running at full capacity. Deliveries keep moving as planned. Throughout the region, we keep executing with highly suitable prudence, robust compliance, and very clear operational controls. What we're monitoring now are the highly practical elements that truly matter. One is logistics routes, along with supplier lead times, local security conditions, FX exposure, collection cycles, and any regulatory shifts that might influence how goods or personnel move. Should anything shift, the effect would probably appear in timing rather than in demand. If that happens, we'll act extremely fast, safeguard delivery, the quality of what we deliver, and update the market once there's something definitive to share. The trends, John, favor us, and they favor us quite strongly, and they're speeding up rather than decelerating. I hope that addresses your question.

John Roy analyst
#52

Indeed it does. In fact, when it comes to trends — and clearly you were discussing AI within India — could you perhaps pull back a bit and walk us through the broader AI landscape, and share what you're seeing take place overall?

Jay Chandan executive
#53

Absolutely — and honestly, that’s a solid question. A lot of folks keep bringing this up to me, and I’ve also discussed it at a number of events. What I’d do is break it down into three trends, as I call them, John. The first one, and these aren’t in any particular order, right? Right now AI is turning into national and regulated infrastructure. When you look at governments, telecom carriers, and regulated businesses, every one of them is viewing AI compute as strategic capacity connected to sovereignty, data residency, compliance, and essential services. Because of that, demand moves away from optional pilots and toward what I’d call funded initiatives that carry long duration and intent. Consider Asia. They’re quickly putting their plans together now and thinking, “We don’t want to be left behind.” They’re now proposing sizable budgets, but even more significantly, they have long duration intent, like I said. The second part of this is the center of gravity, and this matters enormously. I’m not sure why I’m emphasizing it so much, but I’ll emphasize it for you again. The market is getting this completely wrong. People say, “Oh, will the market keep supporting investment in AI?” I mean, companies are putting hundreds of billions of dollars into the U.S. and China. The center of gravity is shifting from training to inference and from inference to distributed inference. Training is extremely lumpy, okay? Inference is extremely persistent. You should take that away. I think most folks on this call — I’m glad for you to take this message. Training is extremely lumpy. Inference, meanwhile, keeps persisting, which means as inference moves into daily workflows, your compute demand spreads across regional hubs and nearer to the data source, which drives more construction of regional data centers. It won’t slow down; it will only rise exponentially. That leads me to my third trend, you know, which is edge. Edge is dramatically widening the addressable market. Edge delivers AI to the decision point, where latency, privacy, and resiliency all count. So what happens now? It speeds up adoption across public safety, as I said earlier, transportation, telecom networks, logistics, industrial operations, and so on and so forth. These things don’t substitute for data centers. It multiplies them. Again, it multiplies them by generating more endpoints that require absolute regional capacity and orchestration. Think of it like this. In the future, you’ll see a lot more of what I call distributed inference points, which will generate an enormous need for regional capacity. That’s why you’ll see players like, you know, OpenAI or Meta or Google or anyone else in the market — they’re shifting across a distributed environment. Those trends favor us extremely strongly, and they’re only speeding up, John; they aren’t slowing down at all.

John Roy analyst
#54

Excellent. Thank you so much for the color.

Jay Chandan executive
#55

Absolutely.

Barrett Boone analyst
#56

Jay and Bruce, nice work on a 2025 that really changed things. My only question is about Quantum-Safe Networks and the SD-WAN offering you have. Could you tell us about specific milestones that those of us investing can watch for?

Jay Chandan executive
#57

Okay. So, like I said earlier, we've built a genuinely powerful product, and over the past several months we've put it through very effective testing. Raja's team has a lot of confidence that they can get it out by the end of April 2026. Now, just so you understand, when we roll out AI infrastructure, it isn't simply a matter of putting GPUs into a room. What we're really talking about is secure connectivity, telemetry, orchestration, and compliance layers, alright? Those things matter enormously. It needs to be clear to everyone that what we sell is not hardware, and we don't rent hardware either. What we deliver is a service. That means your SD-WAN combined with your Quantum Safe encryption lets us govern the network edge all the way to the core with a very high level of security. For us, that raises the value of the solution and it makes the margin mix better. That's point one. Point two is our quantum solutions. Why? Why do people say, "Oh, they're only chasing this because the word quantum is in it"? That's not what we're doing. Anyone who thinks that isn't thinking clearly. These solutions are what make edge AI viable. Why? Because edge compute can only scale when connectivity is intelligent and, even more crucially, secure. What's the role of SD-WAN? What's the role of our post-quantum SD-WAN? It delivers traffic optimization. It enables segmentation along with performance control. Post-quantum crypto future-proofs the transport layer. Once that transport layer is in place, it will help keep that future-proofing intact. Combined with the distributed AI architecture, which we just responded to, it makes these architectures deployable in both a national and an enterprise setting. So what does that turn us into? I believe that's likely where your question was pointing. It stops us from being seen as a rent-or-compute-for-rent type of provider; instead it makes us a trusted operator. That means we're able to design sovereign-grade, Quantum Safe, policy-compliant AI networks. Even more importantly, we can help these GPUs bring in additional revenue and secure the networks that protect it. And even more importantly, our SD-WAN guarantees that neither one collapses as things grow more complex. When the world grows more complex, the way it is right now, we guarantee that our SD-WAN and our quantum won't collapse. Barrett.

Barrett Boone analyst
#58

Thank you. That's very helpful. Congratulations again.

Jay Chandan executive
#59

Thanks, Barrett.

Jay Chandan executive
#60

Much appreciated, Crystal. That genuinely helped. Quite a few sharp questions came my way. A couple took me by surprise too, which is fun. To every investor of ours, to the analysts, and to anyone backing Gorilla, well, let me start by saying thanks. You've placed your confidence in me, in us, and in the whole Gorilla crew for long enough that outcomes, not guesses, now do the talking, right? Certain folks claim our contracts are worthless, that our figures are worthless. Fine. That's just guesswork. The AI infrastructure we're constructing is what governments and vital industries are going to depend on, and our plan is to deliver it with rigor. Anyone familiar with me understands I'm a person who gets things done with rigor. What I'll do now is express gratitude to each of you, then wrap up and pass things along before my tea turns cold. The time is 5:00 A.M., in fact 5:25 A.M., and for me that would be an actual emergency. Thanks, everyone. Enjoy a wonderful rest of your day.

Operator operator
#61

Everyone, that brings today’s conference call to a close. We appreciate your participation, and you are welcome to hang up at this time.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Gorilla Technology Group Inc. transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Gorilla Technology Group Inc. earnings transcripts and 132,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.