OpenAI Partners With Broadcom to Build 10GW of Custom AI Chips

OpenAI and Broadcom announced a multiyear collaboration to co-develop and deploy 10 gigawatts of custom artificial intelligence accelerators and networking systems, marking OpenAI’s first large-scale move into in-house chip design.

Under the partnership, OpenAI will design the accelerators and systems, while Broadcom will develop and deploy racks of AI infrastructure powered by its Ethernet and connectivity technologies. Deployment is scheduled to begin in the second half of 2026 and complete by the end of 2029.

The companies said the collaboration will enable OpenAI to embed insights from its frontier model development directly into hardware, improving efficiency and scalability across its data centers. Broadcom said the agreement reinforces Ethernet’s role as the backbone for next-generation AI datacenters.

“Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential,” said OpenAI CEO Sam Altman. Broadcom CEO Hock Tan called the partnership a “pivotal moment” for advancing toward artificial general intelligence.