Techub News, citing CryptoBriefing, reported that AMP PBC has announced $1.3 billion in financing commitments to build a GPU sharing network. The planned network is designed to aggregate idle computing capacity from independent data centers and provide AI labs with on-demand computing resources. GPUs are widely used for AI training and inference, and the model described by the company focuses on matching computing supply and demand through a shared network.
A computing grid for unused data center capacity
AMP PBC operates as a public benefit company. According to the disclosed information, the company will use a computing grid to connect suppliers of computing power with buyers that need access to those resources. The stated aim is to reduce the cost of renting cloud GPUs. The plan centers on idle resources held by independent data centers, with the computing grid serving as the mechanism for resource coordination and supply-demand matching.
The report also noted that the hourly rental price for a B200 unit has reached $4.89, roughly double the level recorded in January 2026. Against that pricing backdrop, small and medium-sized AI teams are facing difficulties in obtaining computing power. AMP PBC’s financing commitment and GPU sharing network plan are focused on the supply, scheduling and rental cost of high-performance computing resources.