AGPU

• Market Context

  • Axe Compute operates in the global AI infrastructure market, which McKinsey projects will reach $5.2 trillion in total capital spending on AI-capable capacity through 2030. Structurally, demand outpaces available supply in this market.
  • Enterprise customers increasingly seek dedicated, long-term (5- to 10-year) GPU capacity rather than shared generic cloud services, and require a full-service partner to manage equipment acquisition, network design, data center deployment, and ongoing operations — which matches Axe Compute's Build segment offering.

• Contract and Operational Progress

  • Axe Compute achieved $3.2 billion in total signed contract value (TCV) year-to-date through August 2026, after adding $2.9 billion in incremental signed build contracts in the first six weeks of Q3 2026, far exceeding management's prior Q2 projection of $1 billion in new signings for the quarter.
  • The 2K B300 cluster in Columbus, Georgia (the company's first large build project, announced in April 2026) is on track to go live in the coming weeks, and will be tripled in size in subsequent expansion. The company received a $317 million+ prepayment for this expanding cluster.
  • Axe Compute signed a new 55 megawatt multi-location development agreement with partner Duos Technologies, including a joint ownership structure for new data center facilities via a special purpose vehicle that allows Duos to scale development faster with less capital.
  • The company holds $5.9 billion in active, qualified sales pipeline (nearly double the current signed TCV), with two-thirds of pipeline demand for the latest generation NVIDIA Blackwell and Vera Rubin GPUs, and early customer demand already exists for future Vera Rubin architecture ahead of its volume shipment.
  • The company is actively adding more than 20 new employees, all funded by already signed contracts, across deployment, engineering, and customer support teams to deliver on the backlog and pipeline, with a disciplined hiring approach that ties headcount growth to committed revenue.

• Project Financing Model

  • All build contracts use a take-or-pay structure with 20-45% customer prepayment of total project cost up front. After receiving prepayments, Axe Compute pursues off-balance sheet project financing for the remaining project cost, backed by the investment-grade credit of enterprise customers, eliminating the need for incremental equity financing for most projects.

• The company expects to reach an annualized run rate (annualized monthly recurring revenue after full deployment) of $139 million when the April 2026 landmark contract cluster goes live in Q3 2026, and an annualized run rate above $696 million once all currently signed contracts are deployed between Q4 2026 and Q1 2027.
• Management's explicit target is to sign an additional $2 billion in new customer contracts before the end of 2026, and management expects to exceed this target based on current sales momentum.
• Modeled projected margins for build projects are 28-44% gross margin and 62-76% adjusted EBITDA margin.
• Deployed build clusters under currently signed contracts will begin contributing ~$20-$21 million in quarterly revenue after going live, with revenue accelerating as additional clusters are deployed through early 2027.

Axe Compute operates two product segments: Access and Build. In Q2 2026, all $3.2 million in reported revenue (up from $35,000 in Q1 2026) came from the Access segment, which provides on-demand GPU capacity. The Build segment, which designs, deploys, owns and operates dedicated GPU clusters for enterprise customers, has not yet recognized any revenue, as Build revenue is only recognized after clusters go live. A small legacy drug discovery service segment contributed $0.9 million to the quarter's negative adjusted EBITDA, representing less than 100% of total adjusted EBIT loss for the period.

• All signed contracts are subject to deployment execution risk, customer acceptance, and supply chain risks, consistent with disclosures in the company's SEC filings.
• The company holds digital assets (primarily Aether tokens) that are subject to significant fair value volatility; non-cash unrealized fair value changes on these holdings drove almost all of the $17.2 million net loss in Q2 2026.
• Large-scale AI data center deployment is a complex process that requires specialized expertise, and execution delays or operational mismanagement could impact revenue recognition timelines and profitability.

No question and answer section was included in the provided transcript.