NVIDIA Corp.’s new partnership program that introduces a revenue-sharing model with artificial intelligence (AI) cloud operators is expected to grant fast-growing startups and enterprises immediate access to highly sought-after computing power in exchange for a slice of their future cloud revenues.

Under the framework of the partnership programs, NVIDIA would bypass traditional financing bottlenecks that have historically plagued capital-intensive AI infrastructure projects.

Instead of acting solely as a hardware vendor, NVIDIA is positioning itself as an intermediary. The company will collect its standard profit margins on hardware sales to AI cloud operators, while simultaneously securing a recurring, usage-linked stream of the cloud revenue generated by those systems.

The revenue-sharing model reflects a broader industry trend where liquidity-strained AI firms increasingly trade equity or future revenue for compute power. The announcement follows recent reports that NVIDIA is looking to raise at least $20 billion in debt for general corporate purposes, emphasizing the immense capital liquidity required to sustain the ongoing global AI infrastructure boom.

The initial rollout targets key regional hubs outside the United States, foreshadowing NVIDA’s aggressive expansion into sovereign AI clouds rather than relying solely on domestic hyperscalers.

The company named two primary inaugural partners in the Asia-Pacific region, Australia’s Sharon AI and Indonesia’s Firmus Technologies.

Sharon AI is deploying up to 40,000 NVIDIA Grace Blackwell GB300 GPUs across 72 megawatts of new data center capacity. Sharon AI expects to scale its total NVIDIA GPU deployment to over 55,000 units by mid-2027.

Firmus Technologies, in partnership with Singapore-headquartered digital infrastructure platform DayOne, is spearheading a massive DSX AI factory campus in Batam. The 360-megawatt facility is projected to house up to 170,000 NVIDIA GPUs, using advanced Grace-Blackwell, Vera-Rubin, and Vera platforms through 2028. Firmus anticipates between $25 billion and $30 billion in committed offtake agreements during the first six years of the eight-year agreement.

The move highlights a critical structural shift in the tech sector, where graphics processing units (GPUs) have become a scarce, vital commodity that is frequently compared to oil. As the market transitions from early-stage model development to continuous production inference, startups have struggled to secure the traditional capital required to build out massive, always-on data centers.

By restructuring the supply chain to target cloud operators rather than end-users, NVIDIA enables downstream model creators and enterprise clients to bypass site selection, power procurement, and hardware deployment delays.