Nvidia launches revenue-sharing program for AI startups

Nvidia unveiled a revenue-sharing program Thursday that allows AI startups and enterprises to access its computing infrastructure without financing their own hardware buildouts, marking a strategic shift in how the chipmaker monetizes its dominant position in AI computing.

Under the arrangement, Nvidia earns a portion of cloud revenue generated by partners who provide Nvidia-powered services to end customers. The company's CFO Colette Kress described the model as creating a "recurring, usage-linked earnings stream" for Nvidia, according to Bloomberg. For customers, the program eliminates the complexity of site selection, power procurement, construction, and hardware deployment before accessing large-scale compute resources.

The program also includes token credits that enable startups to access compute without upfront capital outlays.

Initial deployments target Asian markets

Nvidia named two launch partners. Sharon AI will deploy up to 40,000 Nvidia Grace Blackwell GB300 GPUs. Firmus Technologies is developing a campus in Batam, Indonesia, planned for up to 170,000 Nvidia GPUs with 360 megawatts of power capacity. The company identified AI-native firms including Baseten, Fireworks AI, and Together AI as target customers for the program.

James Manning, cofounder and CEO of Sharon AI, called the collaboration "a pivotal moment" in delivering sovereign, large-scale AI compute infrastructure. Tim Rosenfield, co-CEO of Firmus Technologies, emphasized that AI-native companies need scalable, energy- and cost-efficient infrastructure to compete globally.

Why it matters

This program represents Nvidia's response to two competitive pressures: the shift from AI model development to production inference workloads requiring continuous operation at scale, and growing competition from custom AI chips developed by cloud giants like Alphabet and Amazon. By offering revenue-sharing arrangements, Nvidia can expand its customer base beyond its largest buyers while maintaining recurring income streams tied to actual usage. The model goes deeper than the equity stakes Nvidia has taken in neocloud operators like CoreWeave and Nebius, creating ongoing financial ties to infrastructure deployment.

Addressing a structural financing gap

Nvidia framed the initiative as addressing a fundamental challenge: emerging AI companies have historically lacked access to capital-intensive compute resources, with even long-term commitments insufficient to secure financing for infrastructure buildouts. The revenue-sharing model shifts the financial burden from startups to Nvidia and its data center partners.

The announcement comes as Nvidia faces intensifying competition from hyperscalers offering their own processors. Nvidia stock was down 0.5% in premarket trading Thursday, according to Barron's.

These details were first reported by Quartz.