U.S. NVIDIA announced on July 1 (local time) a new business model enabling AI-related companies to easily access AI infrastructure equipped with GPUs. It has established a framework combining revenue sharing and financing support with AI cloud providers, supporting the procurement of computing resources by startups, model developers, research institutions, and others.
According to the company, the AI industry is transitioning from a stage centered on model development to one emphasizing production deployment, and demand for computing resources for “AI factories” that continuously generate tokens is expanding rapidly. It explained that this demand requires access to large-scale, multi-tenant accelerated computing that can begin operations quickly and maintain high utilization. Emerging AI companies have historically had limited access to capital-intensive infrastructure, and even when entering long-term contracts, they often failed to secure financing for computing resources.
Under this new business model, AI cloud providers procure NVIDIA infrastructure for AI-native companies, large enterprises, and independent software vendors (ISVs), and provide cloud services. In addition to its customary product sales, NVIDIA receives a portion of the cloud revenue corresponding to the relevant capacity.
The company said the framework will accelerate adoption of its platform in the fast-growing AI-native market while securing a continuing source of revenue linked to usage. It also said model developers, inference providers, agent platforms, and companies advancing the adoption of AI can rapidly access full-stack accelerated computing without having to wait for site selection, power procurement, facility construction, or hardware deployment.
Under the initiative, the company is advancing the construction of AI factories compliant with DSX, NVIDIA’s integrated platform for AI factories. U.S.-based Sharon AI, which operates a sovereign AI business in Australia, plans to deploy up to 40,000 NVIDIA GPUs. Australia’s Firmus Technologies is building a DSX AI factory campus in Indonesia and ultimately expects to expand it to as much as 360 MW, equivalent to approximately 170,000 GPUs.
NVIDIA is positioning the initiative as a new model for procuring AI infrastructure and is also looking to expand the number of participating AI cloud providers.
