OpenAI has signed a landmark deal with Broadcom (AVGO) to develop 10 gigawatts of custom AI accelerators, marking another bold step toward hardware independence.
Under the deal, OpenAI will design proprietary accelerators and systems that Broadcom will help develop and deploy. The racks, scaled entirely with Ethernet and Broadcom’s networking stack, will be installed across OpenAI facilities and partner data centers to meet surging global compute demand.
Says OpenAI co-founder and CEO Sam Altman,
“Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential and deliver real benefits for people and businesses.”
The deal is part of a broader series of chip supply and design agreements struck within a month across the semiconductor industry.
In September, the ChatGPT creator committed $300 billion in spending to Oracle (ORCL), which would provide OpenAI with 4.5 gigawatts of computing capacity as part of its $500 billion Stargate initiative.
Over a week later, OpenAI signed an up to $100 billion deal with Nvidia (NVDA). The deal involves roughly 10 gigawatts of new capacity in Nvidia systems for OpenAI that could power the needs of over eight million US households.
Earlier this month, Advanced Micro Devices (AMD) signed a multi-year agreement to supply artificial intelligence chips to OpenAI. The pact will give the ChatGPT creator the option to purchase up to about 10% of the Nvidia rival, along with access to hundreds of thousands of AMD’s next-generation graphics processors, totaling six gigawatts of compute capacity—enough to power several million US homes.
Disclaimer: Opinions expressed at CapitalAI Daily are not investment advice. Investors should do their own due diligence before making any decisions involving securities, cryptocurrencies, or digital assets. Your transfers and trades are at your own risk, and any losses you may incur are your responsibility. CapitalAI Daily does not recommend the buying or selling of any assets, nor is CapitalAI Daily an investment advisor. See our Editorial Standards and Terms of Use.