OpenAI partners with Broadcom to build 10GW of custom AI chips

OpenAI has announced a landmark partnership with Broadcom to produce 10 gigawatts (GW) worth of custom-designed AI chips, a move aimed at reducing its dependence on Nvidia and strengthening its AI infrastructure.

The collaboration marks a significant step toward OpenAI’s long-term goal of developing superintelligent AI while maintaining control over its own computing resources.

Announced via OpenAI’s official Newsroom account on X (formerly Twitter), the company stated: “We're partnering with Broadcom to deploy 10GW of chips designed by OpenAI. Building our own hardware, in addition to our other partnerships, will help all of us meet the world’s growing demand for AI,” the post reads.

The new chips will allow OpenAI to integrate its frontier model expertise directly into hardware, promising to “unlock new levels of capability and intelligence.”

To put the scale in perspective, the planned output equals the energy production of ten nuclear reactors, signaling one of the largest AI infrastructure projects to date.

According to the company’s announcement, Broadcom will begin deploying the first racks in 2026, with full rollout expected by 2029.

OpenAI CEO Sam Altman called the collaboration “a critical step in building the infrastructure needed to unlock AI’s potential and deliver real benefits for people and businesses.”