OpenAI is collaborating with Broadcom to produce its own AI chips

OpenAI partners with Broadcom to produce its own computer chips to power its AI data centers. The deal is the latest in a series of partnerships designed to reduce the company’s dependence on Nvidia and secure enough computing power to run applications like ChatGPT and Sora and fulfill its mission of developing superior artificial intelligence.

OpenAI said designing its own chips allows it to “embed its learnings from developing leading models and products directly into hardware, unlocking new levels of capability and intelligence.”

Broadcom is expected to begin deploying equipment racks in the second half of 2026, and the deal should close by the end of 2029, the announcement said. Sam Altman, co-founder and CEO of OpenAI, said the partnership “represents a critical step in building the infrastructure needed to unlock the potential of AI and deliver real benefits to individuals and businesses.”

Creating custom chips is part of a growing movement within the technology industry as major players like Meta, Google and Microsoft work to bolster vital supply lines amid rising demand and reduce their reliance on Nvidia’s AI chips. So far, custom chip projects have not posed a credible threat to the chipmaking giant, although they have greatly benefited other chipmakers such as Broadcom.