NASDAQ: AGPU

On September 29, 2025, the Company announced the launch of its Treasury Strategy focused on ATH. Aethir is a leading decentralized physical infrastructure network developed by DCI Foundation, a Panama foundation company (“DCI”), that provides a decentralized GPU network, connecting producers and consumers of GPU compute power at enterprise scale, supporting applications such as artificial intelligence computation, gaming and cloud workloads. ATH functions as a proxy for a unit of GPU compute power and serves as a medium of exchange and unit of incentives for participants in the Aethir network. Participants in the Aethir network can generate yield or other rewards by staking or lending ATH or by otherwise serving as a source of ATH liquidity.

Pursuant to the Treasury Strategy, the Company intends to continue acquiring additional ATH in the open market and to earn yield on its ATH treasury holdings by engaging in ATH staking and other activities. As a holder of ATH, the Company accrues unrealized gains or losses from any appreciation or depreciation, as applicable, in the value of ATH tokens, which trade on various cryptocurrency exchanges.

The Company’s management is focusing its resources on the Treasury Strategy, and a significant portion of the Company’s balance sheet will be allocated to holding ATH pursuant to its Treasury Strategy.

Currently, the Treasury Strategy is primarily dedicated to ATH, and the Company does not intend to allocate treasury assets to other digital assets in the near term. As a result, the Company’s assets will be highly concentrated in a single digital asset. Adverse developments specific to ATH, its protocol, or its network could have a disproportionate impact on the Company’s financial condition and results of operations.

4

The Company’s Treasury Strategy is intended to bring value to its stockholders through the following:

●

utilizing proceeds from equity and debt financings to purchase and hold ATH;

●

staking the majority of the ATH in the Company’s treasury to earn a staking yield and turn the treasury into a productive asset;

●

purchasing locked ATH at a discount to the current spot price; and

●

selling the Company’s ATH holdings, whether on the open market, through block trades, or other negotiated transactions, for various reasons and at various times, including, in order to fund its working capital and general corporate needs.

Axe Compute also continues to operate its Drug Discovery Services business, which we also refer to as our ‘legacy’ business (“Legacy Business”), which is designed to support the discovery and development of optimal cancer therapies. In the Legacy Business, the Company uses AI and its proprietary biobank of 150,000+ tumor samples, categorized by tumor type, to provide actionable insights about drug compounds to improve the drug discovery process and increase the probability of drug compound success. The Company also creates and develops tumor-specific 3D cell culture models mimicking the physiological environment of human tissue, enabling better-informed decision-making during drug development. In February 2026, the Company announced that it is exploring strategic alternatives for this oncology drug discovery solutions business, but the Company’s Board of Directors (the “Board”) has not committed to a specific course of action. Accordingly, the oncology drug discovery solutions business did not meet the criteria under Financial Accounting Standards Board (the “FASB”) Accounting Standards Codification (“ASC”) 205-20, Discontinued Operations to be classified as discontinued operations and held for sale, and therefore is reflected as continuing operations within these consolidated financial statements.

Significant Transactions and Recent Events

On March 20, 2025, the Company completed the sale of assets related to its wholly owned subsidiary, Skyline Medical Inc., to DeRoyal Industries, Inc., a global manufacturer and supplier of medical products. Skyline Medical produced the FDA-cleared STREAMWAY System and associated products for automated medical fluid waste management and patient-to-drain medical fluid disposal and was previously classified as the Company’s Eagan operating segment.

On September 19, 2025, the Company’s stockholders approved a one-for-fifteen (1-for-15) reverse stock split of the Company’s common stock, which became effective at 12:01 a.m. on Tuesday, September 30, 2025.

On October 8, 2025, the Company announced the closing of two previously announced private investment in public equity transactions (“PIPEs”) totaling approximately $343.5 million to support the Company’s adoption of a digital asset treasury strategy focused on ATH, the native utility token of the Aethir ecosystem. The Company raised an aggregate of approximately $343.5 million in the PIPEs from the purchase and sale of (i) an aggregate of approximately 4.4 million shares of common stock (or pre-funded warrants to purchase shares of common stock in lieu thereof) for a purchase price of $11.6265 per share (the “Offering Price”) of common stock (or per pre-funded warrant in lieu thereof) for aggregate cash gross proceeds of approximately $50.8 million (the “Cash PIPE”), and (ii) pre-funded warrants to purchase up to approximately 14.9 million shares of common stock for a purchase price of $11.6165 per pre-funded warrant in exchange for approximately $292.7 million in notional value representing approximately $173.3 million in discounted value of in-kind contributions of locked and unlocked ATH (the “Crypto PIPE”). The pre-funded warrants issued in the Crypto PIPE became exercisable immediately following the Company’s receipt of shareholder approval for the exercise of such pre-funded warrants. The PIPEs closed concurrently on October 7, 2025. The Company has used the cash and in-kind contribution of ATH to fund the Company’s digital asset treasury strategy as well as for working capital and general corporate purposes.

On December 1, 2025, we received formal notice from the Nasdaq Hearings Panel (the “Panel”) of The Nasdaq Stock Market LLC (“Nasdaq”) notifying us that we were in compliance with the stockholders’ equity requirement set forth in Nasdaq Listing Rule 5550(b)(1) (the “Stockholders’ Equity Requirement”). Accordingly, the previously disclosed listing matter has been closed. Nasdaq’s notice further stated that pursuant to Listing Rule 5815(d)(4)(B), the Company will be subject to a mandatory panel monitor for a period of one year from December 1, 2025. If, within that one-year monitoring period, the Nasdaq Listing Qualification Staff (the “Staff”) finds the Company again out of compliance with the Stockholders’ Equity Requirement, notwithstanding Rule 5810(c)(2), the Company will not be permitted to provide the Staff with a plan of compliance with respect to that deficiency and Staff will not be permitted to grant additional time for the Company to regain compliance with respect to that deficiency, nor will the Company be afforded an applicable cure or compliance period pursuant to Nasdaq Listing Rule 5810(c)(3). Instead, the Staff will issue a delist determination letter and the Company will have an opportunity to request a new hearing with the initial Panel or a newly convened hearings panel if the initial Panel is unavailable.

5

On December 11, 2025, we changed our corporate name from Predictive Oncology Inc. to Axe Compute Inc. following the filing of a Certificate of Amendment to our Certificate of Incorporation with the Secretary of State of the State of Delaware (the “Name Change”). Beginning on December 12, 2025, our common stock began trading on the Nasdaq Capital Market under the ticker symbol “AGPU.” In connection with the Name Change, we announced an expansion of our business strategy to include high-performance enterprise artificial intelligence infrastructure. Our historical oncology drug discovery and biomedical research operations, or Legacy Business, includes our AI-driven tumor drug-response prediction platforms, proprietary tumor biobank and related laboratory and biomarker testing services, and were developed under our prior Predictive Oncology business model. We intend to evaluate opportunities for the Legacy Business, including the potential use of the data assets and predictive models developed through those operations in connection with our expanded artificial intelligence infrastructure strategy.

On February 6, 2026, our Board voted to terminate, without cause, the employment of Raymond F. Vennare as Chief Executive Officer, effective February 9, 2026. Mr. Vennare also resigned as Chairman and a member of the Board effective the same date. In connection with his termination, Mr. Vennare entered into a separation agreement providing for severance and other benefits. The Board appointed Chuck Nuzum, an existing member of the Board, as Chairman effective February 9, 2026. In addition, the Board appointed Christopher Miglino as Chief Executive Officer and as a member of the Board effective February 9, 2026. In connection with his appointment, we entered into an employment agreement with Mr. Miglino and granted him stock options to purchase 500,000 shares of our common stock as an inducement award.

On February 24, 2026, we announced that we engaged Cardiff Advisory LLC to assist in exploring strategic alternatives for our Legacy Business. The strategic review process, which is being conducted under the oversight of our Board, may include a potential sale, partnership, licensing arrangement, joint venture or other transaction involving the Company’s biobank platform and related operations. The review reflects our continued focus on advancing our artificial intelligence compute infrastructure strategy while evaluating opportunities to maximize value from non-core legacy assets. There can be no assurance that the strategic review process will result in any transaction.

On March 3, 2026, we announced that our Board appointed Dr. Theodore Zhu and Thorsten Dirks as members of the Board.

Our Business

As of December 31, 2025, we operated in two business areas: (1) the Compute Services and Treasury Management segment, which provides services that include access to GPU compute capacity and manages the Company’s ATH Treasury Strategy; and (2) the Drug Discovery Services segment, which provides services that include the application of AI using its proprietary biobank of 150,000+ tumor samples, as well as creation of proprietary 3D culture models used in drug development.

The Compute Services and Treasury Management segment is the priority of the Company and remains Management’s focus. In regards to the Drug Discovery Services segment, in February 2026, the Company announced that it is exploring strategic alternatives for the business; however, the Company’s Board has not committed to a specific course of action and, as such, does not meet the criteria under FASB ASC 205-20, Discontinued Operations, to be classified as discontinued operations and held for sale. We sometimes also refer to the Drug Discovery Services segment as our “Legacy Business”.

6

Compute Services and Treasury Management

Compute Services

Through our Compute Services segment, we provide enterprises, AI developers, and large-scale workload operators with on-demand access to high-performance GPU compute infrastructure. Our network encompasses global locations and over 435,000 GPUs capable of supporting a broad range of artificial intelligence, machine learning, and high-performance computing workloads. We deliver compute capacity to customers through a managed infrastructure model, typically within 48 hours of customer engagement, without requiring customers to make capital investments in physical hardware or data center facilities.

Our infrastructure is designed to serve both traditional enterprise AI workloads and participants in decentralized physical infrastructure networks. We operate on an asset-light model—we do not own or operate physical data centers—which allows us to scale compute capacity in response to customer demand without material capital expenditure. Revenue from Compute Services is generated primarily through reserved GPU capacity contracts, under which customers commit to capacity on a prepaid basis, typically under 12- to 36 -month terms. Pricing is structured on a per-GPU per-hour basis and is positioned competitively relative to centralized hyperscale cloud providers. We also offer ancillary storage and CPU services billed alongside GPU compute under the same contractual arrangements.

Through the use of ATH tokens as a settlement mechanism, we integrate our Compute Services with our Treasury Strategy. Under this model, Axe Compute may settle our compute obligations using ATH. When we purchase new ATH in the open market, we earn an additional 20% revenue in the form of additional ATH. We believe this integration creates a compounding economic relationship between our compute revenue and our token treasury position that is differentiated from the revenue models of traditional GPU cloud providers.

Treasury Management

On September 29, 2025, we adopted a digital asset Treasury Strategy focused on ATH, the native utility token of the Aethir network. Aethir is a decentralized physical infrastructure network developed by DCI Foundation, a Panama foundation company (“DCI”), that supports GPU compute workloads. ATH is used within the Aethir ecosystem for functions including GPU rentals, staking, validation and network incentives. On December 31, 2025, we held approximately 2.837 billion unlocked ATH with a fair market value of $24.4 million, and a right to receive 3.511 billion locked ATH that are subject to vesting and/or transfer restrictions with a fair market value of $15.5 million after applying a discount for lack of transferability and control. Together, the Company’s treasury included 6.348 billion ATH and future rights thereto which, multiplied by the closing price of $0.0086 per ATH on December 31, 2025, represents a market value of approximately $54.6 million, on an as if vested and claimed basis, that can be deployed to access compute.

The Company intends to continue acquiring additional ATH in the open market and to earn yield on its ATH treasury holdings by engaging in ATH staking and other activities. As a holder of ATH, the Company accrues unrealized gains or losses from any appreciation or depreciation, as applicable, in the value of ATH tokens, which trade on various cryptocurrency exchanges. With the Company focusing its resources on the Treasury Strategy, a significant portion of its balance sheet will be allocated to holding ATH.

Currently, the Treasury Strategy is primarily dedicated to ATH, and the Company does not intend to allocate treasury assets to other digital assets in the near term. As a result, the Company’s assets will be highly concentrated in a single digital asset. Adverse developments specific to ATH, its protocol, or its network could have a disproportionate impact on the Company’s financial condition and results of operations.

The Company’s Treasury Strategy is intended to bring value to its stockholders through the following:

●

utilizing proceeds from equity and debt financings to purchase and hold ATH;

●

staking the majority of the ATH in the Company’s treasury to earn a staking yield and turn the treasury into a productive asset;

●

purchasing locked ATH at a discount to the current spot price; and

●

selling the Company’s ATH holdings, whether on the open market, through block trades, or other negotiated transactions, for various reasons and at various times, including, in order to fund its working capital and general corporate needs.

7

Drug Discovery Services / Legacy Business

Drug Discovery Solutions – PEDAL

Patient-centric Drug Discovery using Active Learning (“PEDAL”™), our proprietary AI-driven platform, offered by our Drug Discovery Services segment, is designed to provide high-confidence drug-response predictions. This platform combines our biobank of samples with a one-of-a-kind database of historical tumor data, and the power of AI to efficiently build predictive models of tumor drug response. Our PEDAL asset is a unique technology that combines one of the largest privately held commercial biobanks of tumor samples, AI active machine learning, and multi-omic historical tumor data – complete with on-site Clinical Laboratory Improvement Amendments (“CLIA”) certified lab testing capabilities to inform drug/tumor model predictions. PEDAL offers researchers the opportunity to incorporate patient diversity early, efficiently, and cost-effectively into the drug discovery process by using data from hundreds of patient samples. PEDAL works by iterative cycles of active learning to guide the testing of patient samples against specific compounds. This results in PEDAL efficiently building comprehensive predictive models of patient drug response in a matter of weeks. This predictive model can rank compounds against tumor samples of certain profiles that respond to specific drugs and can also predict the set of compounds that provide the best coverage across patient tumor samples.

We believe leveraging our unique, historical database of tumor drug responses, genomics, biomarkers, digitized pathology slides, and histopathology data with over 150,000 patient tumor samples to efficiently build AI driven predictive models of tumor drug response will provide actionable insights critical to new drug development. Through the course of over 15 years of clinical testing of patient tumor responses to drugs, our Pittsburgh lab has amassed a huge proprietary knowledgebase of data. To provide for our patient-centric approach, this dataset has been rigorously de-identified and aggregated to inform our proprietary process to create models of tumor drug response.

PEDAL can significantly increase the probability of clinical success by introducing patient diversity early in the development process, while also decreasing the time and cost of oncology drug discovery programs. Our large knowledgebase of tumor drug response and other data, together with proven AI, has created a unique capability for oncology drug discovery, utilizing this highly efficient screening of drug responses against thousands of diverse, well-characterized patient primary tumor samples. With each iteration of a PEDAL campaign, the program learns, predicts, and then directs the most informative wet lab experimentation, while building the predictive model. This allows for a unique and streamlined approach in which AI-driven predictions are tested against samples from this expansive and diverse biobank to more efficiently and effectively narrow down viable drug-tumor pairings. This novel disruptive approach is ideally suited to the early part of drug discovery while also being highly customizable to meet the needs of our collaborators. Our patient-centric drug discovery approach provides for the prioritization of drug compound candidates while accounting for patient tumor diversity. This should dramatically improve the chances of successfully translating discoveries into successful therapies, while simultaneously lowering costs through shortened development timelines, and most importantly, enhanced “speed-to-patient” for new therapies.

A key part of our commercialization strategy has been the understanding that our AI-driven models of tumor drug response serve a key unmet need of pharmaceutical, diagnostic, and biotech industries for actionable multi-omic insights into cancer. In collaboration with these companies, using the predictive models, we will accelerate the search for more effective cancer treatments through biomarker discovery, drug screening, drug repurposing, and ultimately clinical trials with higher probability of success.

PEDAL, which incorporates CORE™, our active machine learning program, with tumor profile data and human tumor samples, provides optimized, efficient, high-confidence drug-response predictions. Our platform is designed to move molecules forward with a higher probability of clinical success. The focus of our business strategy is to leverage and expand our portfolio of proprietary solutions to advance drug discovery and enable oncology drug development for our biopharma partners.

8

3D Modeling

Our Drug Discovery Services segment also develops tumor-specific in vitro models for oncology drug discovery and research. Our 3D tumor-specific models accelerate the drug development process for our clients and partners by providing drug response predictions with high correlation to clinical response, enabling our biopharma clients to manage pipeline prioritization more efficiently.

The 3D models incorporate tissue-specific extracellular matrices and tumor-specific medium supplements allowing for a true reconstruction of tumor microenvironment. Our approach is compatible with multiple classes of immuno-oncology agents from antibody and antibody-drug conjugates to bi- and tri-specific compounds and CAR-T cells. The organ-specific disease models provide 3D reconstruction of human tissues accurately representing each disease state and mimicking drug response.

Our 3D platform maintains tumor-tumor and tumor-stroma interactions and incorporates both cellular and extracellular elements of tissue microenvironment including soluble factors in an organ- and disease-specific manner. It is compatible with multiple cell types, drug classes, and downstream analysis methods. Our models support proliferation of malignant and non-malignant cellular components of tissues.

Applications include providing efficacy screening of anticancer compounds, evaluation of mechanisms of drug resistance, identification of new drug combinations, rescue of failed drug candidates, assessment of off-target toxicity, target discovery and biomarker discovery. Product offerings include preclinical testing services based on our proprietary models directly to clients in the biopharmaceutical industry.

Clinical Testing

Through our wholly owned subsidiary, Helomics Corporation (“Helomics”), reported under our Drug Discovery Services segment, we offer a group of clinically relevant, cancer-related tumor profiling and biomarker tests for gynecological cancers that determine how likely the patient is to respond to various types of available chemotherapy treatments and which therapies might be indicated by relevant tumor biomarkers.

Clinical diagnostic testing is comprised of our Tumor Drug Response Testing (ChemoFx™), Genomic Profiling Testing (BioSpeciFx), and other biomarker tests. The Tumor Drug Response Testing test determines how a patient’s tumor specimen reacts to a panel of various chemotherapy drugs, while the Genomic and biomarker profiling evaluates the expression and/or status of a particular gene or protein related to a patient’s tumor specimen.

Testing involves obtaining tumor tissue during biopsy or surgery, which is then sent to our CLIA certified laboratory using a special collection kit. Tumor Drug Response Testing is a fresh tissue platform that uses the patient’s own live tumor cells to help physicians identify effective treatment options for each gynecologic cancer patient.

Genomic Profiling offers a select group of clinically relevant protein expression and genomic mutation tests associated with drug response and disease prognosis. Physicians can select biomarkers for testing from carefully chosen panels of relevant tests, organized by cancer pathway and tumor type. Results for these tests are presented in a clear, easy to understand format, including summaries of the clinical relevance of each marker.

Industry and Market Background and Analysis

Compute Services and Treasury Management

The global compute infrastructure market represents a significant and growing opportunity. The global cloud computing market— which encompasses the infrastructure-as-a-service and platform-as-a-service segments most relevant to GPU compute— is estimated at approximately $1 trillion in 2026 and is projected to reach approximately $2.4 to $3.3 trillion by 2030 to 2033, representing compound annual growth rates of approximately 16% to 20%, according to multiple independent market research firms. When focused specifically on AI-related computing, worldwide AI spending is forecast to total $2.52 trillion in 2026, a 44% increase year-over-year, according to Gartner. Building AI foundations alone is expected to drive a 49% increase in spending on AI-optimized servers in 2026, representing 17% of total AI spending. McKinsey & Company estimated in 2025 that over 50% of all data center capacity was already dedicated to AI workloads, an amount they expect to grow by a factor of approximately 3.5 times by 2030, and that approximately $6.7 trillion will be spent on data centers globally between 2025 and 2030. Notwithstanding this level of investment, demand continues to outpace supply. As of early 2026, North American data center vacancy rates reached a record low of 1.6% and data center demand increased approximately 24% in 2025, and average GPU procurement lead times stand at 36 to 52 weeks, according to CBRE Investment Management. We believe this supply-demand imbalance creates a significant opportunity for our GPU compute business, which provides distributed GPU compute capacity for AI and other high-performance computing workloads.

9

Drug Discovery Services

The growing demand for the improvement in the discovery and development process of novel drug therapies is driving the demand for AI-empowered solutions. Growing partnerships and cooperation are expected to fuel global market for AI in drug development. The adoption of AI solutions in the drug development process increases efficiency, reduces cycle time, and increases the productivity and accuracy of the risky and long process. Due to these advantages, the importance of AI in drug discovery and development is expected to drive the global market. AI-powered drug discovery is an emerging approach that considers individual variability in multi-omics, including genes, disease and environment to develop effective therapies. This approach predicts more accurately which treatment, dose, and therapeutic regimen could provide the best possible clinical outcome. Biopharmaceutical companies, contract research organizations, academia, and other stakeholders began integrating AI-based solutions in their drug development processes to enhance outcomes and curb costs.

We believe we are uniquely positioned with our PEDAL platform to provide early insights that clients can use to prioritize drugs for development and identify patient-centric indications. In addition, the PEDAL platform can be used to re-purpose previously failed drug compounds. We aim to leverage the PEDAL platform for our biopharma clients and help them prioritize their oncology portfolio. The PEDAL platform supports a biopharma client’s decision on the drug molecules with a higher likelihood of clinical success. With PEDAL, we look to improve/enhance the way that the biopharma industry carries out the development of oncology drugs. We believe our platform provides unique financial- and time-saving advantages for pharmaceutical companies.

We believe the passage of the FDA Modernization Act 2.0 will increase the use of non-animal methods to study the mechanisms of diseases and to test the effectiveness of new drugs. The FDA Modernization Act 2.0 allows for alternatives to animal-testing requirements for the development of drugs and allows drug manufacturers to opt out of animal testing while utilizing other testing methods to develop drugs, such as cell-based assays, organ-on-a-chip technology, computer models, and other human biology-based test methods. We expect the market to continue to grow due to a shift towards more efficient, accurate and predictive models.

Competition and Competitive Advantages

Compute Services and Treasury Management

The market for GPU compute infrastructure and AI cloud services is competitive and rapidly evolving. Our principal competitors include specialized GPU cloud platforms, operators of dedicated AI data center infrastructure, and distributed GPU infrastructure providers. We also compete indirectly with major public hyperscale cloud providers, which offer GPU compute capacity as part of a broader portfolio of cloud services. Many of these competitors have substantially greater financial, technical, sales, and marketing resources than we do, and some have longer operating histories and broader customer relationships in the enterprise market.

We believe our competitive position is supported by several key factors. First, our globally distributed network provides geographic diversity and access to compute resources across multiple regions that we believe is difficult to replicate at speed. Second, our 48-hour provisioning capability and flexible contract structures allow customers to scale GPU capacity in response to near-term workload demands without the multi-month lead times typically associated with dedicated data center deployments. Third, our asset-light operating model—under which we do not own physical data center facilities—allows us to grow our addressable footprint without the capital intensity that characterizes many of our competitors. Fourth, our integration of ATH token settlement creates an additional economic incentive structure that we believe differentiates our offering for customers participating in decentralized infrastructure ecosystems. We believe these factors, together with our pricing model, differentiate us from both hyperscale cloud providers and dedicated AI infrastructure operators. However, there can be no assurance that we will be able to maintain or improve our competitive position, and competitive pressures could materially and adversely affect our revenue, margins, and results of operations.

10

Drug Discovery Services

On average, new oncology drug compounds take 10-12 years to become approved for use, from discovery to commercial launch. Identifying those compounds is a difficult process with a significant majority of compounds failing. This failure is costly in time and resources, particularly when the compounds fail during the clinical trial stages. It is estimated that 90-95% of compounds fail between first human dose and launch. One of the reasons for this high failure rate is the inability of oncology drug compounds in clinical trials to meet the therapeutic end points in a large population.

AI companies addressing the needs in the drug discovery market are looking at the drug discovery and development challenges from different angles. However, we believe no other company has access to a comparable privately held biobank with tumor drug responses, genomics, biomarkers, digitized pathology slides, and histopathology data. The ability to pair AI with our biobank provides us with a competitive advantage and creates a barrier to entry for competitors in the drug response prediction space.

We believe this patient-derived, highly curated, multi-omic tumor model offers a better chance of generating predictive models of drug-response and outcomes than competitive approaches in the market today. The information embodied in the AI-driven predictive model provides insights into each tumor’s response to different therapeutic options, resulting in the ability to provide actionable insights critical to new drug development, individualizing patient treatment, drug repurposing, and biomarker development. Identifying cohorts of patient tumors most responsive to candidate drugs informs the early drug candidate selection process in a patient-centric manner that we do not believe is offered elsewhere. The tumor cohorts identified by our models can also be analyzed and stratified to optimize patient selection criteria for improved clinical trials. A deeper analysis of these same tumor cohorts found to be highly responsive to a particular drug candidate can be further utilized for targeted biomarker development and/or targeted assay development.

We also fulfill unmet needs in the drug discovery market with the next-generation technology of our 3D models, based on extensive knowledge of the human tumor microenvironment creating accurate reconstruction of the organ-specific 3D tissue microenvironment enabling evaluation of therapeutic agents under conditions mimicking human physiology. The main competitive advantage of our technology is the tumor-specific nature of its systems. 3D models replicate tissue heterogeneity and provide maintenance of primary human cells, organoids, and cell lines under the native conditions of human disease. The 3D models are formulated to mimic the tissue and/or disease of interest instead of pursuing a one-size-fits-all approach taken by other companies. Recreating specific tumor microenvironments enables more reliable prediction of tissue response to drugs with varying mechanisms of action. This same technology can also be used to demonstrate potential toxic drug effect on normal tissues by maintaining an accurate reconstruction of cellular and extracellular compartments of human tissues.

Suppliers

Compute Services and Treasury Management

Our Compute Services segment depends on the availability of high-performance GPU hardware and the underlying data center and network infrastructure required to operate it. We source GPU hardware primarily from NVIDIA Corporation, which is currently the dominant supplier of GPUs used for artificial intelligence training and inference workloads. We do not manufacture any hardware ourselves. Our ability to expand compute capacity depends on our ability to procure sufficient quantities of GPU hardware and supporting infrastructure on commercially acceptable terms and within time frames consistent with customer demand.

11

GPU supply has at times been constrained due to strong global demand from AI developers, hyperscale cloud providers, and other large-scale compute operators. We work to maintain relationships with multiple hardware and infrastructure providers across our network of over 200 locations, and our distributed, multi-location model is intended to partially mitigate single-supplier and single-facility concentration risk relative to operators that depend on a smaller number of large data center facilities. However, the loss of or significant disruption to our access to NVIDIA GPU supply, or material price increases or extended delivery lead times from key hardware or infrastructure suppliers, could materially delay our ability to fulfill customer orders, reduce our available capacity, and materially and adversely affect our business, results of operations, and financial condition. We cannot guarantee that we will be able to secure GPU hardware in the quantities or on the timelines required to meet customer demand or execute our growth strategy.

Drug Discovery Services

We buy our raw materials from several suppliers and, except as set forth below, the loss of any one supplier would not materially adversely affect our business. We rely on sole suppliers for certain materials used to perform our molecular diagnostic tests. We also purchase reagents used in our molecular diagnostic tests from sole-source suppliers. While we have developed alternate sourcing strategies for these materials and vendors, we cannot be certain that these strategies will be effective or that the alternative sources will be available in a timely manner. If our current suppliers can no longer provide us with the materials that we need to perform molecular diagnostic tests, if the materials do not meet our quality specifications, or if we cannot obtain acceptable substitute materials, there could be an interruption in molecular diagnostic test processing. In the event of the loss of these suppliers, we could experience delays and interruptions that might adversely affect the financial performance of our business. We have existing and good relationships with our service vendors.

Intellectual Property

So long as our Legacy Business is a part of Axe Compute, we will continue to invest appropriately in intellectual property in order to maintain a competitive advantage in the marketplace. Our Legacy Business relies on a combination of patent, trade secret intellectual property rights, and other measures to protect our intellectual property. We seek to protect our trade secrets and proprietary know-how, in part, with confidentiality agreements with employees, although we cannot be certain that the agreements will not be breached, or that we will have adequate remedies if a breach were to occur.

Government Regulation

Both our Legacy Business and our Treasury Strategy are subject to or impacted by extensive and frequently changing laws and regulations in the United States (at both the federal and state levels) and the other jurisdictions in which we conduct business, including some specific to our business, some specific to our industry, and others relating to conducting business generally (e.g., U.S. Foreign Corrupt Practices Act). We also are subject to inspections and audits by governmental agencies.

Employees and Human Capital Resources

We had 13 full-time employees and 1 part-time employee as of December 31, 2025. None of our employees are subject to a collective bargaining agreement and we believe our relations with our employees are satisfactory. Our human capital resources objectives include identifying, recruiting, retaining, incentivizing and integrating our existing and new employees, and we recruit people for positions regardless of gender, ethnicity or other protected traits.

12

Executive Offices

Our principal executive offices are located at 91 43rd Street, Suite 110, Pittsburgh, Pennsylvania and our telephone number is (412) 432-1500.

Corporate History

We were originally incorporated in Minnesota on April 23, 2002, and reincorporated in Delaware in 2013. We changed our name from Skyline Medical Inc. to Precision Therapeutics Inc. on February 1, 2018, then to Predictive Oncology Inc. on June 13, 2019, and finally to Axe Compute Inc. on December 11, 2025.

Available Information

Our website address is https://axecompute.com/. Information contained on our website is not incorporated by reference into this Annual Report on Form 10-K unless expressly noted.

We file reports with the Securities and Exchange Commission (“SEC”), which we make available on our website free of charge at https://investors.axecompute.com/financial-information. These reports include Annual Reports on Form 10-K, Quarterly Reports on Form 10-Q and Current Reports on Form 8-K, each of which is provided on our website as soon as reasonably practicable after we electronically file such materials with, or furnish them to, the SEC. We also make, or will make, available through our website other reports filed with or furnished to the SEC under the Securities Exchange Act of 1934, as amended, including our proxy statements and reports filed by officers and directors under Section 16(a) of that Act. In addition, the SEC maintains a website (https://www.sec.gov) that contains reports, proxy and information statements, and other information regarding issuers that file electronically with the SEC, including us.

13