Axe Compute Signs $1.3 Billion AI Infrastructure Contracts, Seen as an Undervalued Entry in GPU Computing

From late 2025 to early 2026, as other sectors—such as overvalued growth, cyclical, and purely narrative-driven themes—fade, and AI capital expenditures continue to exceed expectations while semiconductor and data center-related stocks significantly outperform, the market consensus will solidify: “AI is no longer just one theme—it has become the absolute dominant force in global capital markets.” Through our research into undervalued U.S. AI companies, Axe Compute has emerged as a key focus for us this year. The recent announcement on July 22 of a new $1.3 billion AI computing contract has significantly strengthened our confidence in Axe. If this contract is executed smoothly and its impact is accurately reflected in future financial statements, we believe “Axe Compute, currently valued at less than $100 million, could become the most undervalued GPU computing entry point in the U.S. market.”

Before its name change in December 2025, Axe Compute was formerly Predictive Oncology Inc. (NASDAQ: POAI), a typical small-cap biopharmaceutical company listed on the U.S. stock market. As a classic example of a micro-cap biotech stock, POAI delivered modest clinical-stage performance: revenue remained consistently in the hundreds of thousands of dollars, it operated at a continuous loss, and its market capitalization hovered for years in the tens of millions of dollars, drawing minimal attention from capital markets.

In September 2025, the company unexpectedly launched the Strategic Compute Reserve, explicitly centering the Aethir native utility token (ATH) and continuing the Crypto Treasury narrative, signaling a strategic shift toward AI and compute services.

In October 2025, the company completed two concurrent PIPE financings, raising a total of $343.5 million in cash through a hybrid structure consisting of $50.8 million in cash and $292.7 million in nominal value of ATH. This financing transformed the company’s balance sheet from negative equity to a positive $47.7 million in equity, granting it 6.348 billion ATH tokens. The transaction established a deep strategic alignment between the company and the Aethir network, enabling a capital structure that combines AI compute narrative with treasury company dynamics, thereby bringing the company into active investor consideration.

On December 11–12, 2025, the company rebranded, changing its name from Predictive Oncology Inc. to Axe Compute Inc., and its ticker symbol from POAI to AGPU, continuing to trade on Nasdaq.

As the first quarter of 2026 concludes, Axe Compute officially begins operations as a new cloud service provider, with affiliated parties of the crypto project Aethir potentially becoming the largest shareholder, signaling a comprehensive transformation to the financial markets:

On February 9, Charles L. Nuzum assumed the role of Chairman, and Christopher Miglino (previously involved in the ATH transaction structure design) officially became CEO. The board was restructured in March, with Kyle Okamoto (formerly Aethir’s CTO/GM) appointed as President.

On April 1, the company completed its enterprise-grade commercial integration with Aethir’s distributed GPU network (400,000+ GPU containers, 200+ locations, 93 countries), signing its first enterprise contracts totaling approximately $12 million. The contracts are primarily under the Immediate Access Program, generating an expected monthly revenue of approximately $835,000, with payment terms of upfront and monthly prepayments. Minimal compute revenue has already begun to contribute, with approximately $7,000 recognized in Q1.

On April 22, 2026, a $260 million exclusive B300 cluster contract (first order under the Build Program) was disclosed. Key contract terms: a 36-month take-or-pay agreement for delivery of 2,304 NVIDIA B300 GPUs plus high-speed AI storage (in a U.S. Tier-3 data center with 4.8 MW dedicated power). Structured with a deposit, prepayment, and monthly advance payments; quarterly revenue of approximately $21 million is expected after go-live in Q3 2026.

May 27, 2026: Confirmed receipt of the $43 million down payment for the B300 contract, marking the first true cash milestone of the contract and confirming that the Build mode has commenced as planned, with hardware procurement and deployment underway.

June 16, 2026: Execution of a $25.9 million Blackwell/Grace Blackwell long-term deployment contract (12 months + 24 months, renewable), with $12.9 million prepaid.

July 22, 2026: Announcement of new AI infrastructure customer contracts totaling $1.3 billion, based on five-year agreements with renewal options, requiring substantial upfront payments and including provisions for continuous GPU upgrades as next-generation GPUs become available. Revenue is expected to begin generating by the end of Q4 2026, with upfront payments to be made in Q3 2026, at which point annual recurring revenue (ARR) will exceed $384 million. This $1.3 billion order should serve as the true catalyst for the market to reassess Axe.

Axe Compute Inc. is a technology company focused on providing high-performance computing infrastructure for artificial intelligence (AI) workloads. It acquires large-scale GPU capacity from hardware manufacturers and infrastructure providers, then deploys it to enterprise customers through long-term service agreements that include hardware procurement, data center hosting, networking, storage, and financing. Axe also maintains a business unit dedicated to oncology drug discovery solutions, though this is not currently the company’s primary focus.

1. Axe's business is divided into two product lines.

(1) Immediate Access Program

Designed for customers requiring rapid deployment and elastic scaling. Leveraging Aethir’s existing GPU inventory across its distributed network, deployment can be completed in as little as 48 hours, with coverage across more than 200 global nodes. Ideal for inference, fine-tuning, and small-to-medium-scale training scenarios, with monthly billing based on reserved capacity.

(2) Cluster Construction Plan (Build Program / AI Factory)

Designed for massive-scale, long-term dedicated computing needs. Axe handles the overall architecture design, data center location selection and power negotiations, hardware financing arrangements, and ultimately the enterprise-grade SLA (Service Level Agreement) operations—“Design, Deploy, Own, Operate.”

The $260 million three-year contract, set to go live in April 2026, is a landmark case for this model. The company plans to procure a dedicated cluster of 2,304 NVIDIA B300 GPUs and specialized high-speed AI storage infrastructure from a Tier III data center facility in the United States, along with a dedicated 4.8 MW redundant power supply. The customer has specified the deployment location and service standards, with deployment scheduled for completion in the third quarter of 2026. A structured payment arrangement is in place, with an initial payment of $43 million already received. Over the 36-month service period, the company will recognize approximately $21 million in revenue each quarter.

In June 2026, the company signed a $25.9 million long-term deployment contract for Blackwell and Grace Blackwell, covering two use cases: inference infrastructure and simulation platforms. $12.9 million has been received as an advance payment.

In July 2026, the company’s Build business line secured additional five-year AI infrastructure contracts totaling over $1.3 billion across the United States and Europe, significantly exceeding its annual target of $1 billion in contract signings. The advance payments for the project will be received in Q3 2026, with recurring revenue recognition beginning at the end of Q4 2026. Once all clusters are fully deployed and operating stably, the corresponding annual recurring revenue will exceed $384 million. Company management stated that current market demand remains strong, and the related revenue will be reflected in the 2027 annual recurring revenue, continuing to expand the company’s medium- to long-term growth potential.

2. Reconsidering Axe's Build Hashpower Business

The best comparison is Corweave—one is the leader in centralized training, the other is a new global force in hybrid computing power:

CoreWeave follows a capital-intensive, centralized approach focused deeply on training workloads, operating 49 large-scale AI data centers across North America and Europe, with approximately 250,000 high-end GPUs. Leveraging InfiniBand high-speed interconnect networks and native Kubernetes orchestration, it delivers single-data-center clusters of up to ten thousand GPUs, delivering extreme performance for ultra-large-scale distributed training—ideal for trillion-parameter training tasks by leading AI labs such as OpenAI, Meta, and Microsoft. CoreWeave went public on Nasdaq in March 2025 and received an additional $2 billion strategic investment from NVIDIA in January 2026, establishing itself as a benchmark provider in the specialized AI compute cloud (Neo-Cloud) market. However, with all its data centers concentrated in North America and Europe, network latency of 80–150 milliseconds caused by cross-continental data transmission, combined with regional data residency compliance requirements, prevents CoreWeave from serving major markets in Asia-Pacific, the Middle East, Latin America, and other regions.

Axe Compute adopts a hybrid, distributed, and globally covered approach. On one hand, it leverages Aethir’s distributed computing network to integrate third-party data center resources worldwide, deploying over 200 computing nodes across 93 countries and providing access to more than 435,000 GPUs. On the other hand, it is aggressively expanding its asset-centric cloud business, valued at over $1 billion, enabling it to enter the large-scale customized computing market and serve all types of GPU buyers and AI companies.

3. Financial Analysis

Axe Compute's first-quarter financial results as of March 31, 2026

Press Enter or click to view the image in full size.

As of March 31, 2026, the company held $6.9 million in cash and cash equivalents, $20.2 million in ATH digital asset holdings (approximately 2.83 billion tokens), and $9.4 million in current digital asset receivables, totaling approximately $36.5 million in liquidity. Management believes this is sufficient to support the company’s operations through the fiscal year ending 2026 and beyond.

Revenue for the first quarter of 2026 was $35,000, compared to $110,000 for the first quarter of 2025. First-quarter 2026 sales were primarily driven by the traditional drug discovery services segment, with the computing services segment contributing only $7,000. According to company disclosures, a $43 million upfront payment for a B300 bulk order was received in May, and an additional $25.9 million in long-term contracts for the Blackwell series was added in June; neither has yet been recognized as revenue on the income statement.

Once the $260 million dedicated cluster officially launches in Q3, it will generate approximately $21 million in computing power revenue per quarter, equivalent to 600 times the total revenue of Q1. Assuming the $1.3 billion order goes live in Q4, quarterly revenue will increase by an additional $65 million to $86 million, representing a quarter-over-quarter growth of over 400%. The company is on the brink of a breakthrough transition from quarterly revenues in the tens of millions to hundreds of millions, and current market pricing has not fully reflected the certainty of this step-function revenue increase.

Net loss for the first quarter of 2026 was $7.7 million. The net loss included a non-cash, mark-to-market loss of $4.3 million on the company’s ATH digital asset holdings. As of March 31, 2026, accounts receivable amounted to $659,000, compared to $32,000 as of December 31, 2025. Both accounts receivable and contract liabilities increased significantly this quarter, reflecting monthly advance payments due from Compute Services customers following project launches at the end of the first quarter.

Christopher Miglino, CEO of Axe Compute, said: “Our goal this year is to sign contracts worth $1 billion; securing contracts in July has already far surpassed this target… We believe signing an additional $2 billion in contracts this year is well within reach, which will help boost our annual recurring revenue (ARR) next year.” Combined with its public statements in the first half of the year, Axe Compute currently has potential business orders exceeding $4 billion, with over $1 billion in signed contracts, and aims to sign a total of $3 billion in contracts this year.

4. Valuation Analysis

• Model 1: Forward P/S Ratio for FY2026E

Estimated annual revenue

The officially confirmed orders listed below allow us to calculate approximately $125 million in guaranteed revenue for FY2026.

Three Wall Street analysts predict AGPU’s 2026 revenue will average $163,935,524, with a low forecast of $157,505,455 and a high forecast of $168,752,872. In 2027, this figure reaches $254,372,663, with a low revenue forecast of $244,405,017 and a high revenue forecast of $261,853,600. Currently, our conservative estimate based on confirmed revenue is approximately $125 million.

CoreWeave forward PS is approximately 3.88x; Axe Compute's actual revenue for 2026 is estimated at $125 million, with a total share count of 11.385 million, and a current price of $6.85.

Axe market cap = $125 million × 3.88 = $485 million

Equivalent stock price: $485 million ÷ 11.385 million shares ≈ $42.60 per share

Increase multiple relative to current price: 42.60 ÷ 6.85 ≈ 6.21 times

• Model 2: P/ARR (Forward-Looking Scenario Projection)

The P/ARR (Price-to-Annual Recurring Revenue) ratio, which compares market capitalization to annual recurring revenue, is a widely accepted steady-state valuation metric in the compute infrastructure industry. It is particularly well-suited to business models centered on multi-year locked-in compute contracts, as it better reflects the intrinsic value of a company’s long-term, stable cash flows. For this analysis, we reference CoreWeave, the industry leader, whose P/ARR valuation midpoint in July 2026 was approximately 2.4x, serving as a fair valuation benchmark for mature compute service providers.

To date, the company’s total backlog of Build business long-term orders represents an annual recurring revenue (ARR) of $384 million at steady state.

Axe's fair forward market cap = $384 million × 2.4 = $921.6 million

Target stock price = $921.6 million ÷ 11.385 million shares ≈ $80.94 per share

Upside potential relative to current price: 80.94 ÷ 6.85 ≈ 11.8x

Based on the comprehensive valuation analysis, Axe's stock has an upside potential of 6 to 11 times, indicating it is currently significantly undervalued. This valuation does not account for differences in business scale or maturity between the two companies; therefore, the actual fair valuation may be subject to downward adjustment.

From a peer comparison perspective, AGPU’s current market valuation exhibits a significant misalignment with its business scale and growth potential. To date, the company’s market capitalization stands at approximately $80 million, while its guided ARR, based on signed long-term contracts, has reached $384 million, resulting in a P/ARR ratio of just 0.2x. In contrast, peers Nebius, CoreWeave, IREN, and WhiteFiber have P/S on ARR ratios of 6.9x, 2.4x, 4.0x, and 10.4x, respectively. Even accounting for AGPU’s early-stage commercialization and the gradual recognition of revenue, its valuation remains substantially below industry averages. As the dedicated B300 clusters and subsequent contracts exceeding $1 billion in value begin to contribute revenue in the second half of 2026, the company’s ARR is poised to accelerate rapidly next year, offering investors a substantial margin of safety and upside potential given its currently depressed valuation multiples.

In addition to its impressive Compute business, AGPU features a highly scalable flywheel model: the ATH Treasury strategy. Unlike BTC or ETH treasury companies that simply hold coins, ATH is embedded within a company whose related businesses generate positive cash flow. Compute orders directly drive demand for and settlement of ATH, while Treasury appreciation fuels the expansion of Compute. Together, they create a mutually reinforcing, self-amplifying flywheel.

1. What are ATH and Aethir?

The Aethir network is a decentralized physical infrastructure network developed by the Panama-based foundation company DCI Foundation (referred to as “DCI”). The Aethir network aggregates enterprise-grade GPUs contributed by independent data centers, enterprises, and other hardware owners into a globally distributed network. This network is designed to provide instant GPU computing resources for AI training and inference, cloud gaming, and other virtualized workloads, often at lower prices than centralized cloud providers. Within the network, three roles work together to ensure the availability, suitability, and quality of computing resources: “Containers,” which perform the actual computations; “Inspectors,” which test and monitor containers to verify their integrity and performance; and “Indexers,” which match users seeking computing resources with suitable containers. Buyers of computing resources utilize the network’s computing power.

ATH serves as the transaction medium and incentive unit for participants in the Aethir network, acting as a proxy for GPU computational power. To become a computational resource provider, network participants must acquire ATH and stake it as collateral to qualify for contributing GPU resources and processing computational requests. After computational resources are delivered and verified, ATH flows from resource consumers to providers as payment and reward. Resource providers earn ATH through “Proof of Capacity” rewards (for maintaining availability and readiness), “Proof of Delivery” rewards (for completing workloads), and service fees paid by resource consumers. Service providers may re-stake, hold, lend, or sell the ATH they receive. The Aethir treasury manages protocol fees and allocates ATH toward protocol development, while the blockchain settlement layer records transactions and facilitates ATH transfers.

2. Capital Design of Axe and ATH

The capital structure of ATH Treasury is not a simple "buy and hold" model, but rather a two-layer design that deeply integrates Axe Compute’s business entity with the Aethir (ATH) ecosystem, creating a closed loop of business, capital, and token. The core advantage of this design is that every completed Compute order will, in the future, generate incremental demand and value capture for ATH—unlike traditional treasury companies, which rely solely on external market liquidity and sentiment.

(1) Axe's Access business is hosted on the Aethir network.

Access Mode (Instant Access) is the core of AGPU’s asset-light expansion, fully powered by the Aethir distributed GPU network (400,000+ GPU containers, 200+ locations, 93 countries covered). After enterprise customers place orders through AGPU’s Access platform, their compute tasks are directly executed on the Aethir network. Each invocation of an order consumes or stakes ATH, creating real demand.

Each Access order = direct pull on ATH demand + generates positive cash flow (prepaid revenue). This design makes AGPU’s Access business a “natural demand engine for ATH.” Linking business growth to ATH’s price—more orders lead to greater ATH consumption/staking, strengthening price support.

(2) Axe's Treasury Strategy: Hold ATH and Build a Strategic Reserve

AGPU’s treasury strategy is an upgraded version of BTC/ETH treasuries, where companies like MicroStrategy passively hold BTC as “digital gold,” relying on external Bitcoin halvings and market cycles for returns, without intrinsic cash flow generation. In contrast, AGPU’s ATH is integrated within a Compute business that generates positive cash flow. ATH is not merely a reserve—it is staked and settled on the Aethir network, while Axe’s Access services operate atop it, naturally forming a closed-loop system.

3. How does the positive flywheel between AGPU and ATH work?

(1) The business flywheel of AGPU and ATH: An "order-demand-value-add" cycle driven by the Access model

Order fulfillment drives order growth—increased demand for ATH—internal appreciation of ATH—Axe's balance sheet expands—AGPU appreciates—more AI computing power orders are secured.

(2) The Capital Flywheel of AGPU and ATH: A "Performance-Funding-Enhancement" Cycle Driven by Treasury Appreciation

Order fulfillment — Company performance improves, available funds increase — Purchase ATH, increase ATH holdings — External appreciation of ATH — Axe’s balance sheet expands — AGPU appreciates — Secure more AI computing power orders

The business flywheel generates intrinsic demand and cash flow (Access orders directly fuel ATH demand). The capital flywheel provides leverage-driven appreciation and asset expansion (Treasury accumulation amplifies ATH’s price effect), creating two deeply interlocked driving models—performance and ATH price jointly propel AGPU price growth and performance expansion. This model may become a new paradigm for “AI x Crypto” capital frameworks.

The story behind Axe Compute (AGPU) is highly imaginative, and the current market pricing does not reflect optimistic expectations regarding future contract execution and the reserve value of ATH. However, as with any high-elasticity growth stock, the narrative leads financial realization, with valuation based more on future GPU contract deliveries and ATH price performance than on confirmed revenue and profits. Historical revenue remains at a very low base, and the actual conversion of large orders and their validation in financial reports still require time. Below are the key risk factors that investors should carefully evaluate.

1. Contract Execution and Delivery Risk

The Build Program is critical to AGPU’s transition from a light-asset access model to a semi-heavy-asset, customized cluster model. A $260 million B300-exclusive cluster—projected to generate approximately $21 million in quarterly revenue after its Q3 2026 launch—and subsequent $1.3 billion in global customer contracts have already been signed. However, execution risks remain in hardware procurement, data center coordination, power deployment, and the implementation of enterprise-grade SLAs. If the cluster fails to launch on schedule or customer acceptance is delayed, revenue recognition will be postponed, potentially impacting cash flow and market confidence.

2. Revenue Conversion and Financial Statement Verification Risk

Q1 2026 revenue was only $350,000 (with minimal contribution from Compute services), and nearly $1.6 billion in signed orders have yet to be significantly converted into revenue. Wall Street analysts’ average 2026 revenue forecast is approximately $164 million, but these projections assume conversion in the second half of the year. If order fulfillment lags behind expectations, actual revenue could fall substantially below consensus. Non-cash ATH mark-to-market losses will continue to fluctuate, and increases in accounts receivable and contract liabilities also reflect potential bad debt risks under the prepaid model.

3. Macroeconomic and market valuation risks

If AI Capex is reduced due to economic slowdown or technological iteration, order demand may be affected; tighter GPU supply, higher energy costs, and stricter data center compliance requirements could all increase execution costs. Forward P/S and P/ARR calculations are based on the assumption that financial metrics will be realized; the actual fair valuation midpoint may be discounted due to differences in scale and maturity, and the current high elasticity also implies amplified volatility.

Overall, Axe Compute’s narrative has outpaced its financial execution; the timing of revenue recognition will be a key validation point for the next quarter or half-year earnings report. The above risk factors are not exhaustive; investors should conduct their own due diligence, fully understand the associated risks, and make independent decisions based on their own risk tolerance.

Overall, AGPU has successfully transformed from a traditional biotech company into an AI GPU computing provider in less than a year, offering a hybrid AI computing solution that combines a low-capital “Access Model” with large-scale cluster construction and leasing under a “Build Model,” securing an impressive $1.6 billion in order contracts. Combined with its dual-driver “Compute + Treasury” business model, and considering its business structure, asset reserves, and undervalued valuation, AGPU presents a 6–11x upside potential relative to its current stock price, making it a high-elasticity candidate worth close attention in the wave of AI computing assetization.