Axe Compute Secures $1.5 Billion Deal to Deploy Nvidia AI GPU Cluster, Boosting 2026 Contracts

In April 2026, Axe Compute secured a separate $260 million, three-year enterprise deal to deploy 2,304 Nvidia B300 GPUs, expanding Blackwell capacity beyond the flagship agreement.
The company’s Aethir partnership provides access to a distributed GPU network that can spin up capacity in as little as 48 hours.
Expected annual recurring revenue from the newly announced contracts exceeds $384 million, with revenue recognition starting in late Q4 2026.
Management now expects roughly $2 billion in additional contracts before the end of 2026, which would push total 2026 contracted value well above $3 billion.
Axe Compute Inc. (NASDAQ: AGPU) has locked in a five-year, $1.5 billion contract to deploy a dedicated Nvidia Blackwell B300 GPU cluster in the United States, pushing its total 2026 signed contracted value past $3 billion, according to MarketScreener. Shares jumped nearly 12% on the news, MarketWatch reported.
The deal, part of Axe Compute's Build program, covers more than 9,200 Nvidia GPUs for AI training and inference. The company expects over $534 million in upfront prepayments within the next 30 days to help cover hardware and infrastructure costs, TipRanks noted.
The flagship contract will generate more than $384 million in annual recurring revenue, with revenue recognition set to begin in late Q4 2026, according to TipRanks. When combined with earlier Build agreements, the company's annual run rate will hit roughly $696 million. Deployments are expected to start this quarter, with further activations in Q4 2026.
The structure reduces Axe Compute's reliance on equity funding. Instead, the large upfront prepayments will support project-level financing to cover GPU and infrastructure costs, MarketScreener reported. That shift is key for a company still posting ongoing losses.
In April 2026, Axe Compute separately secured a $260 million, three-year enterprise contract to deploy 2,304 Nvidia B300 GPUs, Crypto Briefing reported. That deal expanded the company's Blackwell capacity well beyond the flagship $1.5 billion agreement. Together, the two contracts anchor a fast-growing infrastructure portfolio.
Benzinga noted that the combined deals give Axe Compute a production-scale, single-tenant cluster built entirely on Nvidia's latest Blackwell architecture. That positions the company to serve both large AI training workloads and real-time inference customers.
To move fast, Axe Compute is partnering with Aethir, a distributed GPU network provider. The partnership lets Axe Compute access additional compute capacity in as little as 48 hours, according to TipRanks. That speed matters as AI demand spikes and customers want infrastructure now rather than months from now.
The distributed model also gives Axe Compute flexibility to scale without owning every GPU outright. Management says it now has a strong pipeline and expects to sign roughly $2 billion in additional contracts before the end of 2026, Crypto Briefing reported. That would push total 2026 contracted value well above $3 billion.
Market analysts describe a mixed picture for Axe Compute. The growth numbers are hard to ignore — over $3 billion in signed contracts in a single year. But the company still runs at a loss, and funding large GPU clusters means ongoing financing needs, MarketWatch noted.
Dilution is a concern too. Funding gaps not covered by prepayments could require new share issuance, which would hurt existing investors. For now, management is betting that a fast deployment timeline and a $2 billion pipeline will outrun those risks before year's end, Benzinga reported.
Publishers
10
Articles
23
Reach
33