Axe Compute Secures US$1.5 Bln AI Infrastructure Agreement with NVIDIA Blackwell

Kuala lumpur: Axe Compute Inc, a neocloud artificial intelligence (AI) infrastructure company, has announced a new five-year customer agreement with a total committed contract value of over US$1.5 billion to deploy a dedicated AI infrastructure cluster powered by NVIDIA Blackwell in the United States (US). (US$1=RM4.08)

According to BERNAMA News Agency, the contract, secured through Axe Compute's Build programme, supplies over 9,200 NVIDIA Blackwell B300 graphics processing units (GPUs) in a purpose-built cluster designed, deployed, owned, and operated by Axe Compute.

Axe Compute's Chief Executive Officer, Christopher Miglino, stated that this agreement adds another major contracted deployment to their Axe Build portfolio, bringing the total signed contract value for 2026 to more than US$3 billion. This figure nearly doubles the company's position from earlier this year, and the company aims to maintain this growth trajectory through the end of 2026 and into 2027.

Axe Compute expects to receive over US$534 million in aggregate customer prepayments over the next 30 days in connection with this agreement and the previously announced contracts. These prepayments are anticipated to cover a substantial portion of the associated GPU and infrastructure capital expenditure (capex).

The company plans to combine these prepayments with project-level financing to fund the remaining buildouts. Axe Build programmes are expected to start generating monthly revenue as deployments go online, with initial revenues beginning this quarter and further cluster activations scheduled for the fourth quarter.

Revenue from this agreement, along with previously announced contracts, is anticipated to increase the annual run rate to over US$696 million upon deployment. This represents a significant rise from the US$385 million run rate reported by the company earlier this month.

The contract underscores a customer need for dedicated, single-tenant infrastructure. The Build project is tailored for production-scale AI workloads that require predictable performance, including real-time inference, continuous model fine-tuning, and high-throughput data pipelines.

Through Axe Compute Build, customers can access fully integrated AI infrastructure without having to place data centre assets on their balance sheets or take on the responsibilities of design, deployment, and daily operations.