.svg){kind=spacer}
Details
- NVIDIA and OpenAI have formalized a letter of intent to build out 10 gigawatts of advanced NVIDIA systems for OpenAI's future AI infrastructure, targeting the creation of next-generation models on the path to superintelligence.
- NVIDIA will commit up to $100 billion to OpenAI as each gigawatt of computational power comes online, deploying millions of GPUs across multiple global data centers.
- The project’s first phase, leveraging NVIDIA’s forthcoming Vera Rubin platform, is scheduled to begin operations in the second half of 2026, with hardware and software developed in close coordination between the two companies.
- This agreement extends a decade-long collaboration that began with NVIDIA’s first DGX supercomputer delivery to OpenAI in 2016, underpinning breakthroughs such as ChatGPT and driving an exponential increase in compute capacity.
- The effort will support OpenAI’s 700 million weekly users and complements ongoing partnerships with major players like Microsoft, Oracle, SoftBank, and Stargate partners in building advanced AI infrastructure.
Impact
This $100 billion partnership dramatically raises the bar for AI infrastructure investment, reflecting the enormous resources now required to advance toward superintelligence. The alliance cements NVIDIA's dominance in AI hardware and gives OpenAI the scale needed for continued innovation. The deal is likely to set a precedent for large-scale technology partnerships and shape the competitive landscape for the next era of AI development.
