OpenAI and Broadcom announce massive 10GW AI chip partnership to power next-gen intelligence
OpenAI and Broadcom have officially unveiled their partnership to jointly develop and deploy 10 gigawatts of custom AI accelerators — one of the largest compute collaborations in the industry. The move aims to scale OpenAI’s infrastructure as demand for ChatGPT, Sora, and future frontier models skyrockets. The collaboration leverages Broadcom’s Ethernet stack and custom XPUs to cut costs, boost efficiency, and prepare for a world hungry for faster, cheaper, and more capable AI systems.
1/6

The Deal: OpenAI and Broadcom will co-develop and deploy racks of OpenAI-designed AI chips, set to roll out by late 2026.
2/6

Massive Scale: The partnership involves 10 gigawatts of AI compute power — five times OpenAI’s current 2GW capacity.
3/6

OpenAI’s Expansion: The Broadcom deal adds to recent mega-partnerships with Nvidia, Oracle, and AMD, totalling 33GW of new compute commitments.
4/6

Efficiency Boost: Custom chips built on Broadcom’s Ethernet and memory architecture aim to lower compute costs and energy use.
5/6

Market Reaction: Broadcom’s shares jumped 9% following the announcement, pushing its market cap past $1.5 trillion.
6/6

Bigger Vision: Sam Altman said the goal is “superintelligence at scale,” with faster, cheaper, and more capable models that the world will “absorb super fast.”
Discover the latest Business News, Budget 2025 News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!