Nvidia announced a new financing program aimed at helping AI cloud providers obtain capital for datacenter construction, potentially allowing the chipmaker to earn additional revenue from cloud usage.

AI infrastructure requires substantial investment. Companies such as CoreWeave and Lambda have borrowed billions of dollars from venture capitalists and hedge funds to fund GPU deployments. Their ability to generate profit depends on revenues exceeding loan interest payments.

In a blog post, Nvidia said it will partner with third‑party lenders to sell Nvidia‑powered cloud services, receiving standard product revenue and a share of cloud revenue on supported capacity. The model is intended to accelerate adoption of Nvidia platforms among AI‑focused cloud operators and generate recurring, usage‑linked earnings. Nvidia declined to provide further details beyond the blog post. The company has signed two pilot customers, Sharon AI and Firmus. Sharon AI, a sovereign cloud provider founded in 2024 in Australia, plans to deploy up to 40,000 Grace Blackwell GB300 GPUs. Firmus intends to install up to 170,000 Nvidia GPUs at a 360‑megawatt facility in Batam, Indonesia, built to Nvidia’s DSX specifications.

The program's structure and expected financial impact remain unclear, and it is uncertain how widely other cloud providers will adopt the arrangement.