OpenAI has announced a partnership with Broadcom to produce its first in-house artificial intelligence processors, in a move aimed at securing the vast computing power needed to support its growing services.
The companies said on Monday that OpenAI will design the chips, while Broadcom will develop and deploy them starting in the second half of 2026. The deal involves rolling out custom chips with a combined capacity of 10 gigawatts — equivalent to the power needs of more than eight million U.S. households or five times the output of the Hoover Dam.
Shares of Broadcom rose more than 10% following the announcement. Financial details of the deal were not disclosed, but analysts say the ambitious timeline will require significant funding through strategic investments, pre-orders, and support from Microsoft, one of OpenAI’s key backers.
The partnership comes as major technology firms race to build custom chips to reduce their reliance on Nvidia, which currently dominates the AI accelerator market. Last week, OpenAI also signed a 6-gigawatt chip supply deal with AMD and disclosed that Nvidia plans to invest up to $100 billion in the company to provide data centre systems with at least 10 gigawatts of capacity.
Broadcom and OpenAI expect the deployment of the new systems to be completed by the end of 2029. The project will rely entirely on Broadcom’s Ethernet and networking equipment, challenging Nvidia’s InfiniBand technology.
