On July 2, it was reported that on July 1 (local time), NVIDIA declared the introduction of a groundbreaking AI infrastructure collaboration framework. The primary objective is to collaboratively establish extensive, multi-tenant AI data centers, or “factories,” with AI cloud service providers, leveraging a profit-sharing and credit facilitation approach. This strategic move is tailored to expedite the acquisition of AI computing capabilities for startups, model creators, businesses, and research entities, concurrently creating a sustainable revenue stream for NVIDIA that is directly linked to the utilization of computing resources.

As per the outlined strategy, AI cloud operators will set up AI data centers utilizing NVIDIA's DSX AI Factory blueprint to deliver cloud computing solutions. The initial collaborators encompass Sharon AI and Firmus. Sharon AI intends to integrate up to 40,000 NVIDIA Grace Blackwell GB300 GPUs into its operations. Meanwhile, Firmus is set to construct a DSX AI Factory complex on Batam Island, Indonesia, with aspirations to scale up to a 360-megawatt energy capacity and deploy as many as 170,000 NVIDIA GPUs.