OpenAI partners with Broadcom to develop custom chips and systems for AI accelerators. The deal is for 10 gigawatts of chips, valued at multiple billions of dollars, with deployment starting in the second half of 2029. OpenAI aims to build out 250 gigawatts of compute power by 2033, significantly up from 2GW this year. It would likely cost $10 trillion to achieve this capacity, and current deals have OpenAI on the hook for hundreds of billions of dollars.

In addition to Broadcom, OpenAI has recently struck agreements with NVIDIA, AMD, and Oracle for AI infrastructure and data center capacity. NVIDIA is investing $100 billion, providing 10 gigawatts of AI infrastructure, while AMD is supplying six gigawatts of compute power, with OpenAI potentially taking up to a 10 percent stake in the company. OpenAI is also partnered with Oracle for 4.5 gigawatts in data center capacity as part of the Stargate Project.

Read more at Yahoo Finance: OpenAI is making its own AI chips with Broadcom’s help