Axe Compute (NASDAQ:AGPU) used its fiscal year 2025 investor presentation to outline its transition toward a GPU compute capacity platform, discuss early commercial traction following its rebrand, and walk investors through financial results that management said were heavily influenced by non-cash accounting impacts tied to digital assets and derivatives.
Leadership changes and strategic direction
Christopher Miglino, who said he started as CEO “just over 30 days ago,” reviewed a timeline that included the September 2025 launch of a “strategic compute reserve” and a December move into AI infrastructure alongside a rebrand as Axe Compute on Nasdaq. Miglino also said two new board members joined effective March 26: Theodore Zhu, described as a semiconductor executive with more than 85 U.S. patents, and Thorsten Dirks, described as a European technology executive with prior CEO roles including Telefónica Deutschland and Deutsche Glasfaser.
Commercial update: contract signings and deployment claims
Management emphasized what it characterized as rapid commercial progress in recent weeks. Miglino said the company had executed “approximately $12 million in total contract value over the last month or so,” noting the figure is “subject to deployment and performance terms.” He said Axe Compute was entering Q2 with an expected monthly contract value of about $850,000 as deployments come online, which he said translates to “$7.7 million in contracted 2026 income signed at this time.”
According to Miglino, the company had “more than 20 enterprise customers with over 30 active deployments live.” In closing remarks, he reiterated the contract activity, citing “$12 million worth of contracts” over 30 days and describing monthly contract value as roughly $835,000 across more than 20 customers.
Miglino and Okamoto positioned Axe Compute’s differentiation around three themes:
- Choice: “Any GPU, any location, any configuration” matched to workloads, according to Miglino, supported by a distributed network described as “200-plus global locations.”
- Speed: Deployment “as fast as 24-48 hours, subject to availability,” Miglino said, in contrast to what he described as long wait lists at hyperscalers.
- Distributed delivery: Management emphasized meeting data sovereignty requirements via a multi-location provider approach.
Miglino said contracts are structured with “partial payments in advance,” with customers reserving capacity on a monthly basis. He said the structure is intended to reduce receivable risk and support recurring enterprise income, while covering a range of NVIDIA hardware configurations. He listed examples including H100s, H200s, B200s, RTX 5090, and RTX 5090 clusters.
Strategic compute reserve and token-based model
A central part of the presentation was Axe Compute’s “strategic compute reserve,” which management said consists of tokens used to purchase compute capacity that can then be resold to customers. Miglino said the reserve is composed of an “AI infrastructure token” called ATH and emphasized that token prices are volatile.
Using pricing as of March 27, 2026, Miglino said the company’s market capitalization was approximately $8.6 million, while the strategic compute reserve was valued at $43 million at that point in time. He also referenced an “MNAV ratio,” describing it as a non-GAAP metric, and said it was 0.2 (0.79 on a fully diluted basis). Miglino argued that, at those levels, the public market implied a valuation below the reserve value before attributing value to the operating business or contracted income.
Miglino said the company held 5.9 billion tokens in total, consisting of 2.86 billion “unlocked liquid tokens” and 3.12 billion “locked” tokens that unlock monthly through December 2028. He told investors the company is “not telling investors what the discount should be,” but said management believed the operating business and early commercial traction were not reflected in the market cap.
Okamoto described the operating model as a “flywheel” built around reserved capacity sold to customers with cash paid in advance. He said Axe can also finance and build dedicated infrastructure so customers can remain on an OpEx model rather than deploying capital expenditures. Miglino added that, as capital is used to purchase more ATH, the company “may receive up to a 20% token-based incentive, depending on network mechanics,” which he said is designed to grow the reserve over time.
Okamoto provided an example economics framework, saying a 10-node H100 cluster on a 12-month commitment could generate “around $1.3 million in ARR,” with “gross margin…around 8%, maybe a little higher.” He said customers typically prepay 10% of total contract value at signing and also pay the first month in advance. Okamoto also described potential ancillary services such as shared storage, CPU nodes, and managed Kubernetes as possible margin and “stickiness” drivers.
FY2025 financial results: non-cash impacts dominated net loss
Chief Financial Officer Joshua Blacher said fiscal year 2025 results included “two large non-cash items” that significantly affected the reported net loss. First, he said the company recorded $152.5 million in losses on digital assets at fair value under ASC 350-60, adopted effective Jan. 1, 2025. Blacher said ATH tokens acquired through Dec. 31, 2025 had a cost basis of $102.7 million but declined in market value to $24.4 million at year-end, including a $78.3 million unrealized decline that he characterized as non-cash.
Second, Blacher said the company recorded $52.7 million in losses on derivative instruments related to a “Crypto PIPE” transaction executed Sept. 29, 2025. He said GAAP treatment required a derivative liability at fair value through earnings because the value exchanged involved ATH tokens rather than a fixed dollar amount. He said the derivative liability was derecognized upon settlement on Oct. 7, 2025, when ATH tokens were recorded as assets and pre-funded warrants as equity. Blacher said the $52.7 million reflected the fair value movement between execution and settlement and was “entirely non-cash.”
Blacher said the two items together accounted for $205.2 million of the company’s $233.1 million net loss for the year.
On operations, Blacher reported revenue from continued operations of $125,000 for the fiscal year ended Dec. 31, 2025, up from $85,000 in fiscal 2024. He said the revenue was “entirely attributable” to the legacy drug discovery services segment, and that the compute services segment did not generate revenue in 2025 because the go-to-market pipeline was still being developed during the strategic transition. He said the company expects compute services to become a “meaningful contributor” beginning in fiscal year 2026.
Total operating costs and expenses were $28.6 million in 2025 versus $10.4 million in 2024, Blacher said. Net loss per share was $13.37, based on a weighted average diluted share count of approximately 17.4 million shares, which he said reflected significant pre-funded warrant issuances tied to the PIPE transaction in the fourth quarter of 2025.
On the balance sheet, Blacher said the company ended 2025 with $10.8 million in cash and cash equivalents, compared to $0.6 million at the end of 2024, driven by PIPE financing proceeds in October 2025. He highlighted digital assets of $24.4 million representing 2.8 billion ATH tokens on hand, and a digital asset receivable of $15.5 million representing the contractual right to receive additional locked ATH tokens that are expected to unlock through December 2028. Total assets were $52.9 million at year-end, up from $5.0 million in the prior year.
Blacher said stockholders’ equity was $47.7 million at Dec. 31, 2025, compared to a deficit of $0.2 million at the end of 2024. He added that returning to positive equity and regaining compliance with Nasdaq’s stockholders’ equity requirement in December 2025 were “significant milestones.”
In closing, Miglino said the company’s ability to use its token reserve to acquire compute capacity and resell it for cash could reduce the need to raise additional capital, and he said the company remained “bullish” on its pipeline.
About Predictive Oncology (NASDAQ:POAI)
Predictive Oncology, Inc is a biotechnology company that leverages artificial intelligence and digital biology to support drug discovery and development in oncology. Its core business revolves around the application of machine learning algorithms to high-content cellular imaging, multi-omic profiling, and clinical response data. By integrating these diverse data streams, the company aims to generate predictive models that forecast the efficacy and toxicity of candidate therapeutics, thereby accelerating preclinical decision-making and reducing development timelines.
The company’s primary offerings include its Phenomics platform, which combines automated microscopy with advanced image analysis to capture subtle phenotypic changes in cancer cells.
