Nvidia and OpenAI are embarking on a monumental project to build the world’s most advanced “AI factory,” backed by an investment of up to $100 billion from the chipmaker. This deal, which includes an equity stake for Nvidia in OpenAI, will furnish the AI pioneer with 10 gigawatts of dedicated computing power, an essential resource for developing and deploying “super-intelligence.”
The investment is cleverly structured to align with progress. A foundational $10 billion will be provided once the first gigawatt of computing systems is live. This staged approach ensures that the vast sums of money are tied directly to the successful creation of the infrastructure, which will ultimately be used to train and run AI models of unprecedented scale and capability.
This partnership is the culmination of a decade of collaboration, a point emphasized by Nvidia’s chief executive, Jensen Huang. He noted their shared history of pushing boundaries, from the first DGX supercomputer to ChatGPT. This new, 10-gigawatt initiative is their most ambitious joint project yet, intended to “power the next era of intelligence.”
According to OpenAI’s CEO, Sam Altman, this infrastructure is more than just hardware; it’s the bedrock of the future economy. “Everything starts with compute,” he explained, highlighting that the new systems will be instrumental in achieving future AI breakthroughs. This massive expansion is also a practical necessity to keep up with the intense demand from ChatGPT’s 700 million weekly users.
The first phase is set to utilize Nvidia’s next-generation Vera Rubin platform, with a target launch in the second half of 2026. By making OpenAI a preferred compute partner, Nvidia is cementing its GPUs as the essential hardware of the AI age, while OpenAI secures the resources needed to continue its pioneering work without constraints.
