MUMBAI: Why buy the whole AI factory when you can simply rent the assembly line? Nvidia is betting that the future of artificial intelligence lies not just in selling chips, but in sharing the rewards. The chipmaker has unveiled a revenue-sharing model that aims to lower the financial barriers to AI development while creating a new recurring revenue stream tied to customers’ success.
According to a Bloomberg report, Nvidia will partner with AI cloud providers and data centre operators to give startups, researchers and enterprises access to high-performance computing infrastructure without requiring the hefty upfront investment typically associated with large-scale AI hardware.
Rather than purchasing expensive GPU clusters outright, participating AI companies will gain access to Nvidia-powered infrastructure through the company’s DSX AI factories platform. In return, Nvidia will receive a share of future revenue generated from the computing capacity used by customers.
Nvidia Chief Financial Officer Colette Kress said the initiative is designed to help model builders, inference providers, agentic AI platforms and enterprises scale AI workloads more quickly, bypassing the lengthy process of building data centres, securing power and deploying hardware.
The programme also includes token credits for eligible AI startups, providing additional support for companies developing AI applications under the new partnership model.
The first infrastructure partners include Australian firms Sharon AI and Firmus Technologies. Sharon AI plans to deploy up to 40,000 Nvidia GPUs, while Firmus Technologies is building a new data centre in Batam, Indonesia, designed to scale to 360 megawatts and accommodate as many as 170,000 Nvidia GPUs.
By linking AI developers directly with cloud infrastructure providers, Nvidia hopes to make advanced computing resources more accessible while expanding the reach of its accelerated computing ecosystem beyond traditional hardware sales.
The move also reflects a broader shift in Nvidia’s strategy as competition intensifies in AI infrastructure. Instead of relying solely on one-time chip sales, the company is increasingly positioning itself as a long-term platform provider, generating recurring revenue as customers build, deploy and scale AI applications.
With AI development becoming ever more compute-intensive, Nvidia is wagering that sharing the cost today could secure a larger share of the AI economy tomorrow.
