OpenAI teams up with Broadcom to create custom AI chips
OpenAI is joining forces with Broadcom to design special chips that can handle the massive computing power needed for AI.
These new accelerators, with deployment expected to begin late 2026, could boost OpenAI's capacity significantly.
The announcement even gave Broadcom's stock a nearly 10% jump, showing just how big this move is.
Chips designed with help from OpenAI's own AI models
This partnership has been in the works for a year and a half. By using Broadcom's Ethernet tech and building their own chips, OpenAI hopes to run AI workloads more efficiently while keeping data center costs down.
It also gives them more control over their hardware—something many tech companies are now chasing. Plus, OpenAI has used its own AI models to help design these chips, making the whole process smarter and faster.