Axe Compute Inc (NASDAQ:AGPU) has secured $25.9 million in total contract value across two long-term agreements to deploy Blackwell and Grace Blackwell compute for enterprise AI customers.
The Pittsburgh-based neocloud infrastructure company said $12.9 million of that total has already been received in advance payments, giving it immediate working capital to fund deployment and operations.
Both deals are long-term, running 12 and 24 months respectively, with options to extend.
The two deployments target distinct workloads. One, built on Blackwell GPUs together with compute nodes, large-scale network connectivity and high-speed storage, supports an AI-centric cloud platform delivering inference infrastructure at scale, enabling machine learning teams to train, fine-tune and serve models across a range of generative AI applications.
The second, running on the fully integrated Grace Blackwell GB300 compute stack, powers a simulation platform serving autonomy, gaming and robotics companies, supporting large-scale generation of physics-validated 3D environments, digital twins and synthetic scenarios.
The contracts add to a build-out the company has detailed in recent months, including a previously disclosed $260 million GPU cluster contract on which it reported receiving a $43 million initial payment in May. The latest deals point to continued demand for flexible access to high-end GPU capacity as enterprises expand AI inference and simulation workloads.
Axe Compute, formerly Predictive Oncology, adopted its current name in December 2025 after pivoting to GPU compute infrastructure.