OpenAI reaches 10 gigawatts of contracted AI computing capacity years ahead of target
OpenAI has locked in 10 gigawatts of AI computing capacity, a milestone it once aimed to reach in 2029. The buildout leans on Nvidia $NVDA, which committed up to $100 billion, AMD $AMD supplying six gigawatts of chips, and Broadcom $AVGO on custom accelerators. Total spend is hea
OpenAI is committing to a scale of computing power that would have looked impossible a couple of years ago. The company has now signed contracts for 10 gigawatts of artificial intelligence computing capacity, a level it had originally aimed to reach only by 2029, meaning it hit the goal years ahead of schedule. The buildout depends on a web of hardware partners. Nvidia $NVDA has committed to deploy at least 10 gigawatts of its systems and intends to invest up to $100 billion in OpenAI. AMD $AMD will supply roughly six gigawatts of its Instinct GPUs, with the first gigawatt arriving in the second half of 2026. Broadcom $AVGO is helping design custom accelerators of OpenAI's own. The Stargate infrastructure program, run with Oracle $ORCL and SoftBank $SFTBY, is adding new US data center sites and pushing the total committed spend past $400 billion on the way to a target near $500 billion. For the power grid, the chip industry, and the economics of AI, a single company contracting this much capacity reshapes the entire supply chain.