OpenAI and Broadcom Ink a Monumental 10 Gigawatt AI Chip Deal
OpenAI has announced a strategic partnership with Broadcom to develop and deploy 10 gigawatts of custom AI accelerators. This news sent Broadcom’s stock soaring nearly 10%, adding over $150 billion to its market value.
The collaboration, which has been 18 months in the making, will see OpenAI design custom chips optimized for AI inference. On the other hand, Broadcom will handle the development and deployment. This partnership is part of OpenAI’s broader infrastructure expansion, with deployments expected to begin in late 2026.
“Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential,” said OpenAI CEO Sam Altman. The deal allows OpenAI to embed knowledge from developing frontier AI models directly into hardware, potentially achieving massive efficiency gains and cost reductions.
This announcement follows OpenAI’s recent deal-making spree. The company has formed partnerships with several tech giants, including:
- AMD
- Nvidia
- Oracle
These partnerships total over 33 gigawatts of compute commitments. The custom chip approach could significantly reduce OpenAI’s infrastructure costs. Industry estimates suggest a 1-gigawatt data center costs roughly $50 billion, with $35 billion typically allocated to chips based on current Nvidia pricing.
Broadcom president Charlie Kawwas clarified that OpenAI is not the mystery $10 billion customer the company referenced in September earnings, leaving that client’s identity still unknown.