NVIDIA Launches 'AI Compute Partner Program': Trading Compute Power for Cloud Provider Revenue Share

BlockBeats News, July 2nd - NVIDIA is pushing forward with a new project called the "AI Compute Partnership." According to the cloud service providers participating in the project, Firmus and Sharon AI, NVIDIA has pledged to provide computing power backing for these startup cloud service providers: if the purchased graphics processing units (GPUs) cannot be rented out, NVIDIA will lease back the idle computing power at a guaranteed rate. In return, NVIDIA will take a proportion of these startup cloud service providers' cloud service revenue, with the revenue share proportion gradually decreasing as the contract period progresses.

Due to the significant investment required for computing center construction and GPU procurement, startups with lower credit ratings often struggle to obtain loans. NVIDIA's backing commitment is equivalent to providing a credit endorsement for these enterprises, helping them secure financing smoothly. Currently, giants such as Amazon, Microsoft, SpaceX, Oracle, Meta, and Google have purchased the vast majority of NVIDIA's chips, but these giants are also developing their own AI chips. NVIDIA hopes that by supporting small cloud providers, it can reduce reliance on tech giants.

In addition to revenue sharing, NVIDIA will also receive warrants in some guarantee agreements. For example, NVIDIA previously committed in September 2025 to backstop unsold computing power worth $6.3 billion for CoreWeave until 2032, and in May of this year disclosed an additional $3.5 billion in lease guarantees in exchange for warrants.