Predictive Oncology - Earnings Call - Q3 2025

Executive Summary

• Q3 2025 reported immaterial revenue ($3,618) and a severe net loss driven by a $74.4M non‑cash derivative liability tied to the new ATH token treasury strategy; net loss per share was $(107.25) versus $(7.26) a year ago, and revenue was flat YoY.
• Management pivoted the business to a “Strategic Compute Reserve” on Aethir’s ATH network; as of Nov 10, POAI held ~5.70B ATH (3.7B locked; 2.0B unlocked) with an estimated market value of ~$152.8M at $0.0268 per ATH.
• Cash used in operations improved YTD (nine months) to $5.9M vs $8.0M in the prior year; implied monthly burn averaged ~$0.656M, reflecting cost actions despite minimal revenue generation.
• Catalyst set-up centers on monetizing GPUs via ATH (staking, leasing, token rotation) with near-term enterprise pipeline opportunities and a potential refreshed brand/ticker to align with the strategic focus on AI infrastructure.

What Went Well and What Went Wrong

What Went Well

• Strategic asset accumulation and business model pivot: “We have embarked on a digital asset treasury strategy… Aethir operates the world’s largest decentralized GPU network, offering enterprise-grade AI at a significantly lower cost” (CEO).
• ATH holdings and monetization plan: 5.70B ATH held with a plan to drive “high single-digit yield” in FY26 via staking, GPU leasing, and token rotation (CIO).
• Operational cash burn improving: Operating cash outflow reduced by 26% YTD vs prior year; implied monthly burn ~$0.656M (CFO).

What Went Wrong

• Massive GAAP loss due to derivative liability: $(74.4)M non‑cash loss on derivative instruments tied to the crypto SPA drove net loss to $(77.7)M; EPS collapsed to $(107.25) from $(7.26) YoY.
• Revenue remains de minimis and flat: $3,618 in Q3 vs $3,907 last year; product commercialization remains limited.
• Expense pressure: G&A rose to $2.6M (+$1.1M YoY) on legal and stock-based comp; S&M increased with digital marketing fees despite lower headcount.

Transcript

Operator (participant)

Before we begin, let me quickly remind you that during the course of this presentation, the company will make forward-looking statements. We caution you that any statement that is not a statement of historical fact is a forward-looking statement. This includes remarks about the company's projections, expectations, plans, beliefs, and future performance, all of which constitute forward-looking statements for the purpose of the Safe Harbor Provision under the Private Securities Litigation Reform Act of 1995. These statements are based on judgment and analysis as of the date of this presentation and are subject to numerous important risks and uncertainties that could cause actual results to differ materially from those described in the forward-looking statements.

The company cannot guarantee the accuracy, completeness, or reliability of statements made by third parties or advisors in this presentation, nor can it assure that any expectations, forecasts, or outcomes expressed by third parties or advisors will materialize. The risks and uncertainties associated with the forward-looking statements made in this presentation and webcast are described in the company's public periodic filings with the SEC. Except as required by law, the company assumes no obligation to update these forward-looking statements to reflect future events or actual outcomes and does not intend to do so. The slides presented in this presentation can be accessed via the investors' page of our website.

Raymond Vennare (CEO)

Good morning, everyone, and welcome to Predictive Oncology's Q3 financial results and business update presentation, which will include an overview of our latest digital asset investment, in Aethir's ATH token, aimed at creating the world's first strategic compute reserve. This investment represents a new business line for Predictive Oncology through which we believe we will be able to utilize ATH for revenue-generating purposes. On September 29, 2025, we launched the digital asset treasury focused on Aethir's ATH token, tokens that not only power Aethir's AI infrastructure but also our ability to directly benefit from the active digital asset management of their AI infrastructure. This is both relevant and important because, as we have learned, one of the biggest barriers to accelerating AI-driven medical research and discovery is no longer just about good science but access to reliable and affordable computing power.

A growing constraint on innovation in drug discovery or personalized medicine is not just about the science itself but the infrastructure required to perfect the computational models and develop the AI methodologies that unlock the science. By recognizing this challenge, we believe that we have strategically positioned ourselves with Aethir to address our own future infrastructure needs while investing in an enabling technology that has the potential to unlock the next decade of discovery and innovation, not just in healthcare but across multiple industries. Our plan is to turn Aethir tokens into graphic processing units, or GPUs, via the Aethir partnership and network, and GPUs into revenue. Today's deck will show you exactly how we plan to generate cash flow while supporting the AI compute ecosystem. First, let's start with the agenda so that everyone knows what to expect from today's presentation.

In addition to what I've already said, I will provide you with a brief company update followed by a financial report by our CFO, Josh Blatcher. We will move into our new digital asset treasury business and our vision for the ATH token utilization by Tom McLaughlin, who is our Chief Investment Officer, followed by Kyle Okamoto, a member of Predictive Oncology's Crypto Advisory Board. Kyle will do a deep dive into how we plan to leverage our strategic compute reserve to satisfy expertise, market demand, and generate revenue. As I said in my opening remarks, we were very excited to announce this past September that we have initiated a digital asset treasury strategy focused on the Aethir ecosystem and its native ATH token.

While certainly not the first company to initiate a digital asset treasury strategy, Predictive Oncology's investment in ATH goes well beyond a strategy of simply holding digital currencies in our treasury with the goal of enhancing short-term returns, although this is certainly one very important aspect of the initiative. What really differentiates our digital asset strategy is that by being a participant on the Aethir ecosystem, we can play a direct role in expanding access to high-performance AI infrastructure and significantly lowering costs as compared to the big cloud providers. This represents a new potential revenue stream that we believe will translate into significant long-term value for the company and our shareholders. The growth and demand for AI compute power cannot be overstated, and numerous recent reports have indicated that this demand is rapidly outpacing infrastructure supply.

In fact, one report by Citigroup estimates that AI infrastructure spending will reach $2.8 trillion globally by 2029. We believe our digital asset strategy will allow us to be a beneficiary of this anticipated growth. As a company that has relied on AI to drive our drug and biomarker discovery business, we are well aware of the potential bottleneck that can be created when the availability of AI infrastructure cannot keep pace with global demand or when the cost is prohibitively expensive. We believe this limits breakthrough innovation not just in healthcare but across all sectors. Through this investment in ATH, Predictive Oncology has established the world's first strategic compute reserve, aiming to purchase GPU compute on the Aethir network to help democratize access to AI broadly, while we believe we'll address the rapidly growing need for advanced computing power.

This is an important event in our company's history, and I am excited about the opportunity that this represents. At this point, I will turn the call over to Josh Blatcher, our CFO, for a review of our third-quarter financial performance. Josh.

Josh Blacher (CFO)

Thank you, Raymond. Before digging into the financials, I would like to point out that both our balance sheet and P&L this quarter include some complex and highly unusual effects of the private placement entered into on September 29th to support our new digital asset treasury strategy. They both include transient entries necessitated simply because of the timing of the closing of the transaction. As you all know, in September, the company carried out two private financing deals to support its treasury strategy, one using cash and the other using cryptocurrency, in particular, Aethir tokens or ATH. Both became legally effective on September 29th but closed on October 7th when the company received $50.8 million in cash and $292.7 million worth of ATH tokens, issuing shares and pre-funded warrants in return. The cash deal, or cash SPA, was relatively straightforward.

On September 29th, the company agreed to issue a fixed number of its shares and pre-funded warrants to investors in exchange for a fixed amount of cash from the investors. The only variable that changed between the deal execution on September 29th and closing on October 7th was the company's stock price, which allows for equity classification under US GAAP and no accounting event to be recorded as of September 29th or September 30th. The crypto deal, or the crypto SPA, involved a more complex value exchange. The company agreed to issue a fixed number of pre-funded warrants in exchange for a fixed number of ATH tokens based on an ATH market price determined on September 26th.

The market value of that fixed number of ATH tokens fluctuated between the pricing on September 26th and the close of the third quarter on September 30th and continued to fluctuate until closing on October 7th. As a result, the economic value of the company and investors changed between September 29th and the close of the third quarter on September 30th. Because the economic value changed due to the variable not directly related to the company's own equity, US GAAP requires the contract to be accounted for as a derivative liability at fair value through earnings. The liability was recorded at fair value when the contract took effect on September 29th, then remeasured on September 30th, or the quarter end, and finally again on October 7th at the settlement to reflect the fair value of the underlying ATH contributed and the pre-funded warrants as of that date.

A Monte Carlo simulation was used to model expected price changes in both ATH and the company's stock price over the short period to calculate gains and losses. When the deal closed on October 7, the derivative liability was derecognized, and the ATH tokens were recorded as assets while the pre-funded warrants were recorded as equity. The ultimate impact on the P&L will be an entry to earnings reflecting the ultimate change in the relative market value of what investors contributed, or ATH tokens, versus what they received, pre-funded warrants to purchase the company's common stock. As a result of the foregoing, the financial statements as of September 30, 2025 are not indicative of the company's underlying cash position or financial performance. I will call out a few such entries in this review.

We concluded the third quarter of 2025 with $182,000 in cash and cash equivalents compared to $612,000 as of December 31, 2024. Again, due to the derivative liability of $74 million related to the securities purchase contract supporting the initiation of our digital asset treasury strategy, stockholders' equity of $77 million was compared to the stockholders' deficit of $203,000 as of December 31, 2024. Our net loss for the third quarter of 2025 was $77.7 million as compared to $3.1 million for the third quarter ended September 30, 2024. Again, the significant loss for the third quarter of 2025 was a direct result of the $74.4 million loss on remeasurement of the derivative liability, which was a non-cash charge. Our revenue of $3.6 million in the third quarter ended September 30, 2025, was largely unchanged from revenue of $3.9 million in the comparable period in 2024.

General and administrative expenses of $2.6 million in the three months ended September 30, 2025, was compared to $1.5 million in the comparable period in 2024 and primarily consists of professional and consulting fees. Operating expenses to operations and R&D were $529,000 for the third quarter of 2025 as compared to $535,000 for the corresponding quarter in 2024. Operating expenses to sales and marketing were $133,000 for the third quarter of 2025 as compared to $73,000 for the corresponding quarter in 2024. Importantly, on our statement of cash flows, the cash flow used for operations was $5.9 million for the nine months ended September 30, 2025, versus $8.0 million for the corresponding period in 2024. The reduction of $2.1 million during the period, or 26%, illustrates how significant our cost reduction initiatives have been and how the non-cash loss on remeasurement of the derivative liability skews the P&L.

The monthly burn implied by the statement of cash flows averages to $656,000. This concludes the financial overview. I would now like to turn the call over to Tom McLaughlin, our new Chief Investment Officer, to walk through our investment strategy. Tom.

Tom McLaughlin (CIO)

Thank you, Josh. Hello, everyone. I'm Tom McLaughlin, Chief Investment Officer, and I will walk you through our investment strategy and how we're deploying our assets. Let me start with the composition of our assets and our execution plan. On November 10, 2025, we held approximately 5.7 billion ATH tokens with a market value of approximately $145.9 million US dollars. This figure is calculated based off of a price of $2.68 per ATH token, as reported by the 4:00 P.M. Eastern Standard Time pricing on Coinbase Exchange on November 10, 2025. We've implemented multi-signature requirements for all movement and execution of tokens, with governance and oversight split between Predictive Oncology and our main asset manager, D&A Holdings. We believe this ensures both security and accountability on behalf of all parties involved.

Our investment objective with respect to the digital asset strategy is straightforward: allocate the assets on our balance sheet to drive the best risk-adjusted returns for Predictive Oncology. We are aiming to achieve this objective through a combination of four different allocation models. Our four allocation models: Number one, passive ATH staking in AI and gaming pools. This method provides baseline yield with minimum active management. Number two, active staking for Aethir network cloud hosts. Here, we plan to share rewards from AI bookings participating directly in network revenue generation. Number three, GPU leasing utilizing our ATH. We plan to convert tokens into direct cash flows by booking GPUs and selling compute capacity. Number four, other token rotation across ecosystems.

There's an opportunity to book GPUs using ATH to earn other DePIN, also known as decentralized physical infrastructure network tokens, and either hold those tokens or sell those tokens for cash, which will utilize to purchase additional ATH. The key point here that we'd like to stress is that we consider ATH to be an active asset class and not a passive hold. The goal of this slide is to give a breakdown of the overall assets of the company. As I mentioned in the previous slide, the company currently holds an estimated $5.7 billion ATH tokens on the balance sheet as of November 10th. This equates to an MNAB metric of approximately $0.76. As you can see from the chart on the right, $3.74 billion of these ATH tokens are locked and will unlock through December 2028.

As these tokens unlock, we're able to then leverage them to drive further returns. Our accumulation strategy since the close of the transaction has yielded an additional 927.9 million tokens as of November 10th, utilizing $25 million of cash. We believe Predictive Oncology is well positioned with a large holding of ATH tokens and plans to continue the accumulation of additional tokens off the open market. Summary of objectives. Let me be clear about what we're targeting and why it matters. Our target is to achieve a high single-digit yield on our ATH tokens for fiscal year 2026. We believe this is conservative and achievable given the multiple revenue streams we're activating, and I touched on on the previous slide. We're leveraging different strategies with respect to our digital asset strategy simultaneously: staking, leasing, token rotation to de-risk and diversify revenue sources.

The immediate priority is to drive a baseline of cash flow to sustain our operations. We're not just building for tomorrow; we're generating revenue today. Long term, we're working to create a strong, sustained business model that we believe will be positively valued by the public market compared to competitors, based off of both revenue and profitability metrics. As you'll see when looking at our peer comparisons, the AI infrastructure sector rewards revenue growth and margin discipline, both of which we will be delivering. Overview of Aethir. Now, let me explain why Aethir is central to our strategy at Predictive Oncology. The Aethir network is the foundation of our digital asset strategy, an enterprise-grade decentralized GPU platform serving AI, enterprise, and gaming workloads.

With over 430,000 GPUs deployed across more than 200 global data centers and approximately $150 million in annualized revenue, Aethir is the only decentralized compute platform operating at true enterprise scale. What we think sets Aethir apart is its tokenized participation model. The ATH token is the sole mechanism for engaging with the network, used both to onboard GPU supply and to book compute demand. Every unit of value that flows through Aethir is settled in ATH. In the midst of a global compute shortage, where AI demand outpaces supply by more than 10x, building a strategic treasury in ATH represents a unique opportunity to hold an asset directly tied to the infrastructure layer of the AI economy. Why this matters: AI infrastructure spending among big tech giants is projected to reach $2.8 trillion by 2029.

This isn't a niche market; this is the foundation of the next decade of technological growth. The network model that Aethir brings represents a new way of doing business. We are excited about the avenues this system provides. Kyle Okamoto will now double-click on the utility of the strategic compute reserve and how we plan to convert it into enterprise revenue.

Kyle Okamoto (CTO and AI General Manager)

Thank you, Tom. Good day, everyone. I'm Kyle Okamoto, member of Predictive Oncology's Crypto Advisory Board and CTO and AI General Manager of Aethir. Today, I'll walk you through the overall AI infrastructure market context and, most importantly, how Predictive Oncology will leverage their strategic compute reserve to generate real-world utility and the next steps to return shareholder value as a result. Now that Tom is giving you a brief overview of the Aethir network, let's zoom out a bit and take a look at the overall global AI infrastructure market. This data really tells a compelling story. You know, Tom mentioned that Citigroup projects that AI infrastructure spending will hit $2.8 trillion by 2029. The same forecast has only gotten bigger as each quarter goes by. By 2023, the global power capacity demand is projected to be 200 gigawatts, according to Bain & Company.

While buildouts are happening today, they are still in the single-gigawatt ballpark, so very much early days. The AI infrastructure market is growing at approximately 40% CAGR through 2030. This isn't linear growth; this is exponential acceleration. There's a clear supply and demand imbalance of over 1,000%. Let me repeat that number. The demand for compute exceeds supply by more than 10 times, right? Jensen from NVIDIA, Elon from Tesla and X, Sam from OpenAI, and more importantly, our enterprise clients that we work with every day all attest to the lack of access to affordable and appropriate AI infrastructure. The top five hyperscalers alone will spend nearly half a trillion dollars on AI CapEx by 2026, according to both Citigroup and Reuters. This is inherently the problem with such a precious resource being dominated by a small group of companies.

Aethir's mission, combined with Predictive Oncology, is to change that, right? Predictive Oncology is now in a position to help fulfill that mission. We do not believe this is just a temporary blip. The well-versed analysts, research firm, big bank, creditor, and investor communities have continually updated their forecasts around demand for GPU compute, and we believe this trend will only continue in the years ahead. To Raymond's earlier point about structural constraints, an average of 40-50-week lead times for enterprise GPUs means that companies cannot wait to buy hardware themselves or wait for traditional infrastructure providers to ramp up. They need solutions now. I want to share a quote from Sam Altman recently, CEO of OpenAI, where he stresses that building a strategic national reserve of computing power makes a lot of sense. That is exactly what Predictive Oncology has done: building a strategic compute reserve.

The strategy is quite simple, yet powerful. Let's double-click on that. It really starts with the establishment of the strategic compute reserve, which is now done and announced, and we're now looking forward. Predictive Oncology has built the industry-standard core component to unlock new ways of deploying compute, and we have a great team in place that's already starting to put that reserve to work. Step two is to stimulate supply. This is in flight now. Predictive Oncology is aiming to monetize and grow their ATH holdings via staking GPU clusters on Aethir's network, providing loans to cloud hosts that contribute GPUs to the network, stimulating supply, rewarding the community ultimately with better options and more options, while further satisfying that insatiable demand and earning accretive value for doing so. Step three is to satisfy that demand, which is also in progress.

Predictive Oncology anticipates converting ATH to fiat revenue by renting GPUs on the Aethir network and then providing that AI infrastructure to enterprises directly. We will double-click on this point in a minute. Lastly, step four is to grow the reserve. Predictive Oncology anticipates buying more tokens on the open market with cash proceeds and receiving an additional 20% of the number of ATH tokens purchased by Predictive Oncology on the open market from the Aethir Foundation. Also, by generating ATH rewards for staking and lending to physically grow the overall ATH war chest. That creates a virtuous cycle of satisfying enterprise demand and growing not only the digital assets on the balance sheet, but cash holdings as well.

In short, we're converting capital into productive compute at market speed, which we believe will help solve that 40-52 week supply chain problem while growing the strategic compute reserve and generating positive cash flow. Okay, we can't have a presentation without the infamous eye chart. Here's that chart showing a detailed view of the transaction workflow that generates the main utility, not the only utility, but the main utility of the native ATH token, which is satisfying enterprise demand for premium AI infrastructure. Let's walk through this and how it works in practice. It's a simple four-step operation. Step one is Predictive Oncology lends ATH to cloud hosts for staking GPUs, and they then earn additional ATH rewards on the Aethir network. Step two is Predictive Oncology utilizes ATH to rent GPUs to serve enterprise clients.

Predictive Oncology in step three then earns positive margin fiat cash revenue, generating cash flow from delivering those GPUs to those enterprise clients. Step four is Predictive Oncology uses a portion of that fiat revenue to purchase more ATH on the open market and receive a 20% grant from the Aethir Foundation, with the remaining cash supporting operations, further investments, and profits. This creates a virtuous cycle. ATH generates compute access, compute generates cash, cash generates more ATH, and more ATH generates more compute capacity. This capital allocation engine is designed for continuous optimization across staking, rental, and sales cycles, all while aligning financial returns with market expansion objectives.

This also allows any cloud host or GPU provider to engage with the Aethir network, as Predictive Oncology can support them via staking, lending, and currency appropriation, further stimulating the supply side of the equation, delighting those enterprise clients with more choices, faster solutions, and better prices. That brings us to the near-term pipeline and how we're monetizing this demand today. Predictive Oncology, in partnership with Aethir, has access to a near-term pipeline with enterprise-scale revenue potential. Let me walk you through a few small samples of some of the significant opportunities we're currently developing. These six active prospects span GenAI platforms, inferencing leaders, multimodal AI companies, et cetera. These aren't speculative. These are well-funded, high-growth companies, mostly based in the U.S., backed by top-tier VCs and strategic investors like NVIDIA, AMD, Meta, Sequoia, et cetera.

You'll notice that the contract sizes range from about $3 million for a short-term training run for a real-time generative video firm on NVIDIA H200 GPUs. This is a company that's already successfully used Aethir in the past for such training runs. It ranges up to a $120 million long-term engagement on the newest, latest, and greatest NVIDIA GPUs, the Blackwell B300 series, which is being moved to production globally this quarter. Beyond these six examples, there's 100-plus more prospects in the pipeline, ranging from startups that want to experiment with GPUs for the very first time on the brink of their next groundbreaking innovation, to universities or research firms that desperately need GPUs at the right price to complete their studies and come out with that next big idea.

While the latest, you know, AI unicorns in the enterprise space may get the limelight on this slide and this earnings call, and the latest and greatest Hopper and Blackwell GPUs are obviously highlighted here, the go-to-market teams are satisfying a wide customer base across multiple GPU configurations. As AI is very much horizontal, it's touched or is touching all industries, all markets, all geographies, all verticals. Therefore, the team aims to have a diversified client portfolio, very much in line with the vision and mission to democratize access to such a precious resource. I would say the key takeaways are high-end, industry-leading AI infrastructure is in high demand at the right price. There are several major enterprise opportunities being developed in partnership with Aethir, with numerous small deals progressing. Contract durations are usually long-term, right? This is reserved, dedicated bare-metal infrastructure.

We shoot for six months or longer, and payment is always in advance, to lower risks. The market we now fully operate within represents billions in unmet demand that Raymond described. We're really positioned to capture a meaningful share. Let's take a look at what that might mean from an investor perspective. AI infrastructure peers trade at 19-67x, you know, P to S ratios despite losses. Predictive Oncology aims to deliver profitable revenue with zero infrastructure, hardware infrastructure requirements, creating significant scarcity value. To put this in market context, let's look at the NeoCloud public company peer group and highlight the key differentiation points, because what we're doing is rare and unique in this space. When you look at the AI infrastructure sector, investors strongly favor revenues and growth, largely ignoring large losses.

Companies like CoreWeave, White Fiber, Applied Digital, and Nebius are trading at PS multiples ranging from 18, 19 to 67x, despite massive net losses of hundreds of millions of dollars per quarter. Their stock prices have clearly skyrocketed as investors know the AI narrative is here to stay and we are in the early days. Here's what's different. Predictive Oncology aims to have profitability on each AI infrastructure customer deal that it may execute on the Aethir network. While NeoCloud competitors are burning cash to build infrastructure, Predictive Oncology can provide large GPU clusters with zero physical infrastructure investment needed. That's the power of the strategic compute reserve. It's instant access to AI infrastructure on an OpEx basis, utilizing the native ATH token to unlock this capacity.

Similarly, this peer group has significant operating costs and overhead, where Predictive Oncology gets a ready-made turnkey solution for GPUs as a service without the need for massive software development budgets. This allows Predictive Oncology to remain highly price competitive in the space and ultimately make GPUs affordable for developers working on that next life-changing innovation for researchers looking to tackle life's biggest challenges. I'm showing you the numbers as of a few weeks ago in the NeoCloud peer group, October 15. What's really unique here is that no public NeoCloud that is generating positive net income and free cash flow exists. That puts Predictive Oncology a cut above the rest. While the market is rewarding growth and revenue, not profitability, Predictive Oncology ultimately aims to deliver both revenue growth and profitability. We're quite excited about what's coming soon.

Predictive Oncology is expanding beyond medicine to become a key player in AI infrastructure. As we look ahead to what is next on the near-term horizon, beyond closing deals in the later stages of the pipeline, as I just previously discussed, Predictive Oncology is really anchoring the mission around compute infrastructure powering the new AI economy. We believe that the Predictive Oncology strategic compute reserve built on Aethir's ATH token will allow Predictive Oncology to use ATH to provide the latest and greatest AI infrastructure to enterprises globally. That creates real utility for the digital assets in that reserve, generating revenue and profit for the company to further grow the reserve and ultimately delight clients. This massive war chest of compute tokens is expected to allow Predictive Oncology to onboard NVIDIA's latest GPUs, those Blackwell B300s, to satisfy marquee enterprise logos that crave that bleeding-edge technology to push innovation forward.

Lastly, the team anticipates a near-term refreshed brand and stock ticker to reflect the strategic focus, continuing Predictive Oncology's leadership in AI and machine learning-based drug discovery and therapy techniques, while adding what we hope to be a high-growth business around AI infrastructure and digital asset management, with more exciting digital asset utility use cases to follow. This includes this portion of the prezo, and we will go back to Raymond for closing remarks. Thank you.

Raymond Vennare (CEO)

Thank you very much for your attention, and this marks the end of today's presentation. We remain very optimistic about the future and exceptionally enthusiastic about our partnership with Aethir and D&A.