Key Takeaways
- OpenAI and Broadcom partner on 10-gigawatt custom AI chip project
- Deployment of new AI accelerators begins H2 2026, completes by 2029
- Part of OpenAI’s broader strategy to secure massive compute capacity
OpenAI has signed a major multi-year partnership with semiconductor company Broadcom to develop and deploy 10 gigawatts of custom AI accelerators. This massive infrastructure project, set to begin in late 2026 and finish by 2029, aims to meet the exploding global demand for artificial intelligence capabilities.
The custom chips will be designed by OpenAI and manufactured in collaboration with Broadcom, utilizing Ethernet and other Broadcom connectivity solutions. These systems will be deployed across OpenAI and partner data centers worldwide.
According to OpenAI, designing its own hardware allows the company to directly implement lessons learned from developing frontier AI models, potentially unlocking new levels of capability and intelligence.
Leadership Perspectives
OpenAI CEO Sam Altman emphasized the strategic importance of the partnership: “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.”
He added, “Developing our own accelerators adds to the broader ecosystem of partners all building the capacity required to push the frontier of AI to provide benefits to all humanity.”
Broadcom CEO Hock Tan commented: “OpenAI has been in the forefront of the AI revolution since the ChatGPT moment, and we are thrilled to co-develop and deploy 10 gigawatts of next generation accelerators and network systems to pave the way for the future of AI.”
OpenAI’s Expanding Compute Partnerships
With over 800 million weekly active users, OpenAI has been aggressively securing compute capacity through multiple high-profile partnerships. Last month, Nvidia committed up to $100 billion to deploy 10 gigawatts of AI data center infrastructure, including OpenAI’s deployment of Nvidia’s future Vera Rubin chips starting in 2026.
Additionally, just last week, OpenAI signed a multi-year deal with AMD to deploy billions of dollars worth of AMD GPUs. This agreement includes six gigawatts of AMD GPUs, beginning with one-gigawatt of AMD Instinct MI450 GPUs in the second half of 2026.
