Image source: The Motley Fool.

Date

May 18, 2026

Call participants

• Chief Executive Officer — Christopher Miglino
• Chief Financial Officer (outgoing) — Joshua Blacher
• Chief Financial Officer (incoming) — Jeremy Yaukey-Witter
• President — Kyle Okamoto

Takeaways

• Net loss -- $7.7 million, with $4.3 million of that attributed to noncash mark-to-market losses on digital assets due to Aethir token price declines.
• Total revenue -- $35,000, down from $110,000 year over year; driven by reduced legacy Drug Discovery Services sales and minimal new Compute Services segment contribution.
• Operating loss -- $3.4 million, separating out the noncash digital asset component.
• Contract liabilities -- Increased to $786,000 from $144,000 sequentially; approximately $650,000 relates to Compute Services customer prepayments for revenue to be recognized in future periods.
• Digital asset holdings -- Fair value of $20.2 million, down from $24.4 million at year-end, reflecting both a decline in token prices and mark-to-market accounting.
• Digital asset receivable -- $15.4 million in future Aethir token vesting, with $9.4 million classified as current and $5.9 million as noncurrent; vesting extends through December 2028.
• Cash and cash equivalents -- $6.9 million as of period end, falling by $3.9 million primarily due to operational spending.
• Cash used in operating activities -- $3.7 million, significantly higher than the prior year’s $1 million, largely due to increased working capital and professional services outlays.
• Total assets -- $45.2 million, with total liabilities of $5 million and stockholders’ equity of $40.3 million, down from $47.7 million at year-end, primarily due to the quarterly net loss.
• Largest contract in company history -- $260 million, thirty-six-month agreement for 2,304 NVIDIA B300 GPUs at 4.8 MW in a U.S. Tier 3 data center, signed in April; deployment targeted for the third quarter of 2026.
• Anticipated revenue from landmark contract -- Expected at approximately $21 million per quarter for the duration of the thirty-six-month term, contingent on cluster deployment timing.
• Customer prepayments -- Model includes 15%-30% upfront payment, with monthly advance billing throughout the term and hardware remaining as a company-owned asset.
• Pipeline highlights -- $4.3 billion total contract value across forty-five qualified prospects representing more than thirty-six thousand GPUs; 72% of pipeline requests are for Blackwell family GPUs, with most contracts trending toward thirty-six- to sixty-month terms.
• Revenue model -- Compute contracts are recognized ratably as services are performed; company emphasizes shift to recurring revenue and asset accumulation via equipment ownership.
• Management transition -- CFO responsibilities have shifted from Joshua Blacher (outgoing) to Jeremy Yaukey-Witter (incoming), reflecting a move toward a full-time, dedicated finance function.
• Name change and NASDAQ listing -- Company rebranded as Axe Compute (AGPU 5.64%) and began trading under AGPU in December 2025.
• Strategic compute reserve -- “Tokens to buy compute” held on balance sheet, viewed by management as “additional cash” for liquidity, though not classified as such under GAAP.

Need a quote from a Motley Fool analyst? Email [email protected]

Risks

• Digital asset exposure -- CFO Yaukey-Witter stated, "Our reported net loss for Q1 2026 was $7.7 million. I want to explain that number upfront because much of it is noncash in nature and driven by accounting rules that require us to mark our digital asset holdings to market every quarter." This highlights sensitivity to Aethir token price declines.
• Declining legacy revenue -- CFO Yaukey-Witter said, "The year-over-year decrease reflects reduced sales in our legacy Drug Discovery Services business, which remains in continuing operations as we evaluate strategic alternatives."
• Elevated cash burn -- CFO Yaukey-Witter reported cash used in operating activities of $3.7 million, up from $1 million prior year, linked to staffing, severance, and professional services. This resulted in a $3.9 million decrease in cash during the quarter.

Summary

Axe Compute (AGPU 5.64%) outlined its transformation into an AI compute infrastructure provider, emphasized by the signing of a $260 million, thirty-six-month contract for more than two thousand three hundred NVIDIA B300 GPUs with initial financial impact targeted after deployment in the third quarter of 2026. Management’s commentary described a recurring revenue model built on customer prepayments and company-ownership of deployed hardware, supported by a compute pipeline totaling $4.3 billion in qualified prospects, most seeking multi-year agreements. The company’s balance sheet is significantly influenced by digital asset holdings and receivables, subject to regulatory mark-to-market adjustments that drove a substantial noncash loss in the reported period. Operating results demonstrated declining legacy revenues offset only marginally by early compute segment contribution. Cash burn accelerated during the quarter due to strategic business transition and associated expenses.

• CEO Miglino said, "once this thing gets installed, we'll be seeing $21 million every 3 months." This clarifies the expected cadence of revenue recognition from the landmark contract.
• President Okamoto provided that more than thirty-six thousand GPUs are in the pipeline, with 72% demand for Blackwell-class chips, indicating pipeline concentration in leading-edge hardware.
• Management reported that contract liabilities increased approximately 5.5 times sequentially, driven by new customer prepayments on compute agreements.
• CFO Yaukey-Witter said, "The company's cash used in investing and financing activities was insignificant during the first quarter," highlighting that operational spending, not capital deployment, drove cash decrease.
• CEO Miglino referred to the ability to convert compute reserve tokens into cash as needed and relatively quickly, although GAAP does not recognize these as cash equivalents.

Industry glossary

• Compute reserve: Proprietary pool of tokens held by the company, used to purchase or access data center compute resources as needed.
• Blackwell (GB200, GB300, B200, B300): NVIDIA’s series of high-end GPU architectures, frequently requested for large-scale AI and compute workloads.
• N+1 redundancy: Data center power configuration providing an extra (“plus one”) backup component to ensure uninterrupted power supply for critical hardware.
• Contract liabilities: Deferred revenue on the balance sheet representing advance or prepayments by customers for which services/contracts have yet to be performed or recognized under GAAP.
• OpEx model: Operational expenditure structure where costs are paid over time through usage, rather than upfront capital investment (CapEx).
• SPV: Special purpose vehicle, a legal entity created for the purpose of isolating financial risk in specific transactions or projects.

Full Conference Call Transcript

Christopher Miglino: Thank you, Erin. Good morning, and thanks for joining us. I'm Chris Miglino, the CEO of Axe Compute. With me today are Josh Blacher, our outgoing Chief Financial Officer; and Jeremy Yaukey-Witter, our incoming CFO; and Kyle Okamoto, our President. Today we're presenting our Q1 2026 earnings. We appreciate you taking the time to hear about the company. I realize that a lot of you may be new to our story. So today, I want to tell you 3 things: what we do, why it works and where we're going. Enterprises have been told to work with the constraints of whatever compute data centers happen to have available: on their timeline, at their price, in their location.

If you want GPUs, they tell you to get in line, 36 to 52-week wait list, and you have to take the region that the data center offers. You have to take the architecture they support. They want the off-takers, the renters of that equipment, to configure their business around what they can give you. That's the compromises that enterprises have been making in AI compute. So we see it differently. We're partnering with businesses to build on their terms. Three things make that real. Choice. So any kind of GPU, any location, any configuration match to the workload that the company needs.

And we are seeing customers wanting all kinds of GPUs, not just the latest GPUs that are out there. There is a gamut of customers that want stuff that has been out there for years. Enterprise. We handle sourcing, matching, logistics, so our customers build around their businesses. And then we have global reach. So our supply team is engaged with talking to data centers globally, that allows us to deliver inventory that might be available very quickly, and then other customers that may want to build custom build-outs at these different centers all over the globe.

So we're experts at going out and finding the power that's necessary that is available for the off-takers to do what they need to do. The time lines at the bottom shows how fast we've executed on our plan. At the end of September, we launched the Strategic Compute Reserve, enabling Axe to own tokens to buy compute. The best way to understand this is the same way consumers and teams could use tokens to engage with OpenAI and Claude, we use our tokens to buy bare metal compute. That turns into cash for the company. That's why investors should view our compute reserve as additional cash.

While it's not technically viewed that way from a GAAP perspective, we can turn those tokens into cash as needed and relatively quickly. We then rebranded and listed as AGPU on NASDAQ in December. In February, I started as the CEO. Even though it's new, role as a CEO, I've been involved in structuring this entire transaction with the legacy public company, and helping negotiate the terms with the legacy company and helping do the raise for the initial creation of the compute reserve. So I'm well aware and I was up to speed on the company. In February, we then closed $12 million in compute transactions. And then we closed a $260 million landmark deal in April.

We've assembled an amazing team that sells to off-takers, the people who rent the compute, and the team that sells to the data centers. I couldn't be happier with the team that we've assembled right now. Everything is working out great with the team. So let me tell you a little bit about this large transaction we did, because I know that we announced the transaction in a press release, but we've never talked about it. First off, it's the largest contract that the company has done in our history, albeit that we got going here since February of this year, but this is a foreshadowing of what we think is going to be the future for the company.

But the contract is a 36-month agreement for 2,304 NVIDIA B300s at 4.8 megawatts of dedicated N+1 redundant power in a U.S. Tier 3 data center. The targeted deployment is Q3 of 2026. So this is all in progress right now. The customer pays monthly in advance regardless of their utilization on a committed basis. There's -- obviously, the power and the redundancy is there inside the data center. We signed and announced this deal at the end of April, but the build is actively underway. The machines have been ordered, the power has been allocated, the hardware is being deployed. The first financial installment is expected to be received this month, and it's very substantial.

The revenue recognition for this transaction when the cluster goes live translates to around $21 million per quarter in revenue. That's done. That's once this thing gets installed, we'll be seeing $21 million every 3 months. So if it launches in month 2 of the next quarter, we'll recognize 2/3 of that. But $21 million per quarter going forward for the next 36 months. We think that this model is what's going to drive value for Axe and our cash flow and equity value. So let me show you a little bit about how that works. I want you to understand why enterprises choose Axe to partner with.

The customer begins with their workload size and regional location requirements to us. They bring them to us. Axe then does 3 things. We match the right GPUs, the right data center location, and all the supporting infrastructure: networking, power, rack, NVIDIA Reference Architecture and so on. Then we help arrange the financing of the build-out. In some cases, there's SPVs that are created. In others, we fully participate in the equity and the debt in the transaction. The consumers pay monthly in advance, usually with a 15% to 30% prepayment upfront, and the balance billed monthly in advance. And at the end of it all, Axe owns the hardware that we provided the equity for.

The equipment sits on our balance sheet as a hard asset. When the customer's term ends, we are then able to redeploy that equipment to the next customer. So we're able to utilize the capital to finance the equipment. Think of it the way institutional investors think about data center real estate buildings. They amortize them over 30 years. We're doing the same with GPU clusters over 3, sometimes 5. Every deal we close builds asset value on the balance sheet, not just cash flow. Even though these transactions are cash flow positive, they're eating down the debt that we put on to the equipment that we put in place when we deploy the clusters.

As we grow, we can be seen as a virtual data center owner, that doesn't own the buildings, but we own the equipment globally. We've seen there is still strong demand for NVIDIA chips that came out 3 years ago. In fact, some older chips are being rented out now for more than they had 2 years ago. So the value of these chips that are amortizing off over the next 3 years will be significant at the end of that time frame, is our prediction. So with that said, I'll now hand the call over to Josh Blacher, our outgoing CFO. Josh has been an important part of the company's journey.

I'd like to thank Josh, who has acted as a fractional CFO for the company for some time now. The company is at a point where we're ready for a full-time, dedicated CFO. And while we'll miss Josh and appreciate his contribution, we're excited to welcome Jeremy as the new CFO of the company. Josh?

Joshua Blacher: Thank you, Chris. I've had the pleasure of serving the company as Chief Financial Officer since September 2023. As previously announced, this will be my last earnings call as I transition out of this role. I'm excited to introduce my successor, Jeremy Yaukey-Witter. Jeremy, congratulations on this new role. The company is in excellent hands. Jeremy will walk you through the financials for our first quarter ended March 31, 2026. Jeremy?

Jeremy Yaukey-Witter: Thank you, Josh. I'd first like to take the opportunity to thank Josh for his service to the company. We appreciate the foundation he's helped establish and which I'm excited to build on. And to everyone listening, it's a pleasure to join you today. I started with the company 3 years ago and previously served as the company's Controller. The foundation of my career was built at KPMG, where I provided audit and attestation services to publicly traded and private companies across various industries, including technology and energy. I've been along for each step of our strategic transition, and I'm thrilled about the direction in which the company is headed.

Before I go through the numbers for our first quarter ended March 31, 2026, I want to frame what you're looking at in our filings. Because our financial statements for the first quarter reflect the transitional state of our business at that time with 2 different segments, at very different stages, they also contain certain noncash items that I want to make sure are clearly understood. Looking at our income statement. Our reported net loss for Q1 2026 was $7.7 million. I want to explain that number upfront because much of it is noncash in nature and driven by accounting rules that require us to mark our digital asset holdings to market every quarter.

Specifically, we recorded $4.3 million in losses on digital assets during the quarter. Under U.S. GAAP, our Aethir token holdings are carried at fair value at each reporting date, with changes flowing through the income statement. The price of the Aethir token declined during the first quarter, which resulted in a $4.3 million noncash mark-to-market loss on our token holdings. Setting aside that digital asset fair value adjustment, our underlying operating loss was approximately $3.4 million, reflecting our operating cost structure as we continue to transition the business to focus on our Compute Services segment. Total revenue for Q1 2026 was $35,000, compared to $110,000 in Q1 2025.

The year-over-year decrease reflects reduced sales in our legacy Drug Discovery Services business, which remains in continuing operations as we evaluate strategic alternatives. Our Compute Services segment contributed minimal revenue in Q1 2026 related to just a handful of compute contracts that commenced at the end of March. In line with U.S. GAAP, we recognize revenue on our compute contracts ratably over the service period. Our contract liabilities, representing customer prepayments received ahead of revenue recognition, increased from $144,000 at December 31, 2025, to $786,000 at March 31, 2026. Approximately $650,000 of that balance was related to Compute Services and represents contracted revenue that will be recognized as compute services are provided in subsequent periods.

Total operating costs and expenses were $3.5 million in Q1 2026, compared to $2.4 million in Q1 2025. Q1 2026 expenses were primarily driven by general and administrative expenses of $2.9 million, up approximately $1.1 million from Q1 2025. The primary driver was a onetime recognition of severance expense to our former CEO following his departure at February and the Board's appointment of Chris to the CEO role, along with other personnel related costs. On a per share basis, the net loss was $0.36 per share based on a weighted average share count of approximately 21.2 million shares. In accordance with U.S.

GAAP, that weighted average share count included the 14.7 million prefunded warrants, which remained outstanding as of March 31, 2026, down from the 16.8 million prefunded warrants that were originally issued pursuant to our October 2025 private placements. We'll now turn to cash flows and the balance sheet. The company's cash and cash equivalents decreased by approximately $3.9 million during the quarter, from $10.8 million at December 31, 2025 to $6.9 million as of March 31, 2026. The decline reflects our operational spending during the quarter. Cash used in operating activities was $3.7 million in Q1 2026, compared to approximately $1 million in Q1 2025. The increase primarily reflects increased cash used in working capital and increased cash operating expenses.

Cash used in working capital primarily reflected payments of outstanding accounts payable and accrued expenses. Increased cash operating expenses primarily reflected cash payments for additional professional services resulting from the company's adoption of its treasury strategy in late 2025. The company's cash used in investing and financing activities was insignificant during the first quarter while the company focused on transitioning its operations. Our digital asset holdings had a fair value of $20.2 million as of March 31, 2026. That compares to $24.4 million at December 31, 2025. The decrease primarily reflects the mark-to-market adjustment I described earlier, directly tied to the decrease in the market price of the Aethir token during Q1.

Our digital asset receivable had a fair value of $15.4 million as of March 31, 2026, representing our contractual right to receive additional Aethir tokens in future periods pursuant to time-based vesting conditions, also referred to as locked Aethir tokens. These locked tokens vest on a predictable schedule through December 2028. Of that total, $9.4 million was classified as current, representing tokens expected to vest and be claimed within the next 12 months, and $5.9 million was classified as noncurrent. Something I'd like to once again draw your attention to is the company's accounts receivable and contract liabilities balances, which you'll see each jumped by more than $600,000 from December 31 to March 31.

This jump reflects billings for noncancelable and nonrefundable monthly prepayments due from our compute customers. These billings reflect the traction that the company saw at the end of the first quarter as we began successfully closing contracts with customers. Total assets as of March 31, 2026 were $45.2 million, compared to $52.9 million at December 31, 2025. Total liabilities were $5 million, and total stockholders' equity was $40.3 million, down from $47.7 million at the prior year-end, with the change driven primarily by the $7.7 million net loss for the quarter. I'll now turn the call over to Kyle Okamoto, our President, to discuss our business efforts.

Kyle Okamoto: Thanks, Jeremy. Six weeks into the job now and I will say that I am even more fired up and excited about this market and this company's future. So today, I'm going to cover 3 things. One is how our business model has evolved to support our customers. Second is the Axe build delivery model in depth. And then the third is our pipeline, where the numbers are powerful to say the least, and I think they tell a very compelling story of where this business is going. All right. So the first is how customers grow with Axe. We really have 2 tracks for engaging with customers.

The access available-now compute and we will build out a dedicated cluster, right? So path one and path two. Most customers enter through Axe Compute's immediate access inventory. It's fast, they leverage our existing compute network, they can start in as little as 24 to 48 hours. These clients come to Axe Compute because they have an immediate need and they can't find what they need through traditional hyperscalers or even other neoclouds. The triggers are very consistent. It's not just wait lists of 36 to 52 weeks. It's also the type of inventory they can't find elsewhere, the locations that they need to meet their own customer expectations, like data sovereignty and inference performance.

And of course, the pricing transparency that all clients in this world should expect. And then the level of service that they ultimately deserve. Track two is our build program, that Chris talked about with the $260 million landmark deal. These customers are looking for large and dedicated infrastructure coverage in a specific footprint. They want to partner for the entire build-out, both technically and financially. They come to us for our expertise and ability to find the space and power, finance and source the chips and other equipment, manage the build-out and operations end-to-end. These engagements are typically 3 or 5 years, with cost efficiency as a priority, and of course, enterprise-grade SLA requirements throughout.

Across both of these tracks or paths, customers value the same things, right? They want choice. They want to be able to get what they want when they want it and how they want it. They need expertise and support. And ultimately, global reach, because the world is a very flat place nowadays and you need to be where your customers are, you need to support data sovereignty, and you ultimately need to have very high-performance requirements. And the fourth is really our service.

So our job is to make sure that regardless of how clients want to engage, clients never have to go anywhere else, whether they stay on track one or move to track two, the experience, the support and the relationship are seamless. Now let me walk you through what it actually means for Axe Compute to deliver a large cluster like the $260 million deal. Because some may question, why would somebody choose Axe Compute for something this big? The answer is that we have a full infrastructure partner relationship. We are not a vendor here. And here's what it really looks like across these 6 steps.

From an architecture perspective, we design the full solution with their needs at the center. That could be what type of GPUs, what type of CPUs, what type of onboard storage, high-speed shared storage, what network fabric choices, private and public networking, power and cooling, rack layouts, ancillary services, take Kubernetes or Slurm as an example. It's really all in the aim to build the true AI factory for our clients. The customers tell us what they need and we figure out how to get the job done. Second is financing and procuring. So Axe funds the build. We source the hardware, NVIDIA B300s, Grace Blackwell GB300s, whatever the job requires.

We deal with orchestrating lead times on various components, all with transparency so clients know what they want to know throughout the process. The customers do not have to bring their own CapEx and we're providing this on an OpEx model. We, of course, secure data center space and power in the right location, making sure that we have right of first refusal to support expansions for our customers, making sure that we provide N+1 redundancy to support enterprise-grade SLAs or higher, really all match the data sovereignty and performance requirements. And then, of course, we've got to build, deliver and support, which is happening now.

So we install, test, hand over, provide response times, resolution times, SLA, ticketing system, integrations, et cetera, of course, using NVIDIA Reference Architecture. And we make sure that the client is very much involved in that process from the get-go. Another unique point of Axe Compute is that we offer an upgrade path. So customers can move to the latest GPU architecture during the term. So we don't want to lock them in on an older GPU generation and not allow them to move to the latest and greatest, to keep up with the market innovation and growth, with, of course, Axe Compute supporting that growth. And at the end of the day, Axe Compute then owns the hardware, right?

At the end of an Axe build, the equipment is on our balance sheet as a hard asset. We own it, we can redeploy it, we can provide GPUs as a service to different clients. Every deal builds that asset base. And that last point is really what makes this business model attractive beyond just positive cash flows, right? It's repeatable and it compounds. Now let me talk about the pipeline. Before I show you the pipeline -- anyways, let me set the market context here, right, because these numbers matter. We're operating in a market that people like McKinsey and Gartner estimate has already crossed $1 trillion in AI compute in 2026. That's with a T, trillion.

And it's growing at a 30-plus percent CAGR through 2034. Total data center CapEx is projected to be over $6.7 trillion through 2030, from McKinsey. So obviously, the demand is there, and it's showing no signs of slowing down. It's quite the opposite actually. So now here's what that demand looks like in our pipeline after just a few months of operating, as Chris mentioned, really since starting this business earlier this year. And of course, after closing our first large Axe build deal in my first 3 weeks at the company, that obviously accelerates quite a bit on the pipeline, right? And I want to be clear here, this -- a pipeline is a pipeline, right?

These are qualified prospects, not signed contracts. Not all of them will close. We share this because we believe it's material context for understanding the opportunity in front of us, not as a forward-looking commitment to any specific revenue outcome, right? This is a sampling of, I think, there's 45 prospects in this more advanced pipeline stage. That's representing about over 36,000 GPUs. Of that mix of GPUs, 72%, so about 26,000 GPUs, are Blackwell requests, Blackwell, Grace Blackwell, B200, B300, GB200s, GB300s, et cetera. And that really shows that our clients want the most powerful compute in the world right now for their cutting-edge use cases and innovations that are ultimately reshaping and touching every aspect of the world.

And this is a pipeline, total contract value across these qualified deals of over $4.3 billion. What's interesting here is that the commitment length mix is very telling. The majority of these deals trend toward 36 and 60-month commitments. Customers understand that it's very important to secure critical AI infrastructure to power their businesses, and they are willing to commit to that to ensure that they have a secure and steady supply of compute to power their businesses. These are not spot requests for a few hours or a few days. Enterprises are making these long-term infrastructure decisions and looking for a true partner.

So the bottom line here, if we close, let's say, 3 or 4 of these deals this year, that's hundreds of millions of dollars in incremental contracted revenue on top of the $260 million already signed. So we're quite excited, and that's definitely one of our key focuses. So with that, I will hand it back to Chris for closing remarks.

Christopher Miglino: Thanks, Kyle. Kyle and the team have been doing an amazing job. We're setting the foundation for long-term success at the company. We're at the right place at the right time more so than anything I've ever seen in my career. The deals we're making now will make the company for years to come. I think that the transactions that will get done this year will provide revenue that will be substantial over the next 3 to 5 years. And we don't see that slowing down at all. So our approach is it's scaling fast. We're not the data center, but we're the company that builds and owns what's inside the data center.

I know we've covered this point a lot, but the first deal that we've done has opened a floodgate of opportunities for us. In addition to the $21 million a quarter that particular transaction will bring, it's led to a pipeline of over $4.3 billion, if you missed that in what Kyle said earlier. I've publicly said that I believe that we close around $1 billion in transactions this year. And seeing the pipeline numbers that are with the team, I'm very confident that will happen. So off-takers, the people that rent the equipment, like our model and they want to work with us. And the most important part is building asset value.

It's the part that matters the most for investors thinking about long-term value. Our 2 tracks add meaningful revenue assets to our balance sheet. With Axe build, every cluster we deploy adds owned AI infrastructure to our balance sheet, assets that amortize over the term and redeploy to the next customers. We're not just generating cash flow, we're building the company with a strong foundation for the future. So with that said, I want to thank you for your time and your consideration and interest in Axe Compute.

I think that, obviously, the numbers that are reflected in this first quarter are not representative of the business itself, and you'll start to see that kick in the second and third quarters of this year when the model that we've put in place truly starts to reflect to the income statement cash flow and to the balance sheet. So thanks for being with us here today, and we appreciate your time.