Key Points
- OpenAI is partnering with Broadcom to create its first in-house AI chips.
- OpenAI will design the chips, and Broadcom will develop and manufacture them.
- The deal is for a massive 10 gigawatts of custom chips, to be deployed starting in 2026. The move is designed to reduce OpenAI’s reliance on Nvidia.
- This is the latest in a series of huge chip deals for OpenAI, following recent agreements with AMD and Nvidia.
OpenAI is teaming up with chipmaker Broadcom to produce its first-ever in-house artificial intelligence processors. The move is the latest in a series of massive chip deals for the ChatGPT creator as it scrambles to secure the immense computing power needed to fuel its AI ambitions.
The news sent Broadcom’s stock soaring more than 10%. Under the new partnership, OpenAI will design the chips, and Broadcom will develop and deploy them, with the rollout starting in the second half of 2026. The scale of the project is staggering: they plan to deploy 10 gigawatts worth of custom chips, which is enough to power over 8 million homes.
This is a major strategic move for OpenAI. By designing its own chips, the company is looking to reduce its dependence on costly and supply-constrained processors from the market leader, Nvidia. It’s a strategy that other tech giants like Google and Amazon have also pursued.
The deal is just the latest in a flurry of activity from OpenAI. Last week, it announced a 6-gigawatt chip deal with AMD and revealed that Nvidia is planning to invest up to $100 billion in the company.
While analysts don’t expect this new in-house chip to challenge Nvidia’s dominance immediately, it’s a clear sign that OpenAI is building a powerful and diverse supply chain to ensure it has the resources it needs to stay at the top of the AI world.
