Nvidia will invest up to $100 billion in OpenAI, and OpenAI will build and deploy at least 10 gigawatts of AI data centers with Nvidia systems.
OpenAI and Nvidia announced a strategic partnership and plans to deploy at least 10 gigawatts of AI data centers with Nvidia systems for OpenAI’s next-generation AI infrastructure.
The first phase is expected to come online in the second half of 2026 and will include Nvidia’s Vera Rubin combination CPU-GPU platform to train and run new models. To support this deployment, including new data center hardware and a total of 10GW of power capacity, Nvidia plans to invest up to $100 billion in OpenAI.
[ Related: More Nvidia news and insights ]
Some aspects of the deal will be sure to raise eyebrows among Nvidia’s customers and competitors alike. OpenAI will work with Nvidia as a preferred strategic compute and networking partner for its AI factory growth plans. In addition, the two companies have committed to “co-optimize” their respective roadmaps for OpenAI’s model and infrastructure software and Nvidia’s hardware and software. This could lead to OpenAI having a bigger say in Nvidia’s plans than before.
A Nvidia spokesperson said that this deal is separate from Project Stargate, the $500 billion data center project announced earlier this year featuring OpenAI, Oracle, and SoftBank. It launched with much hoopla but has since struggled to gain traction.
OpenAI is already an exclusive AI partner for Microsoft, offering ChatGPT through the Bing search engine and Microsoft Office 365. Microsoft promised in January to invest $85 billion in AI data centers. However, that deal seems to be unraveling. OpenAI Has partnered with Oracle to offer its services through Oracle Cloud Infrastructure, while Microsoft has added Anthropic’s Perplexity generative AI service alongside ChatGPT.
OpenAI’s next-generation data centers will use Nvidia’s Vera Rubin platform, which went into production in August and is expected to begin shipping late next year. The Vera Rubin NVL144 is expected to offer 3.6 exaflops of FP4 performance and 1.2 exaflops of FP8 performance.
But that performance comes with a cost: power. GPUs are incredibly power hungry. To put 10 gigawatts of power into perspective, a typical nuclear power plant generates about one GW of power. And OpenAI and Nvidia have plans to consume 10 times that amount. It’s no wonder, then, that the data center industry is looking to nuclear power plants to charge their data center ambitions.
