Catenaa, Wednesday, June 24, 2026- OpenAI announced its first custom chip developed with Broadcom called Jalapeño, which provides better performance per watt than current state-of-the-art chips.
The AI startup claimed the new chip was built for inferencing with its own AI models, as well as those across the industry. Inferencing is the process of running AI models.
“Jalapeño is part of our long-term full-stack infrastructure strategy to make compute more abundant, resulting in AI which is faster, more reliable, more affordable for people and businesses, and can be used to solve more important problems,” OpenAI president Greg Brockman said in a statement.
“By designing more of the stack ourselves, we can serve more intelligence with greater efficiency and keep pushing advanced AI toward broader access.”
Broadcom stock rose by about 1.5% on Wednesday. The stock is up by over 11% so far this year.
OpenAI said it took just nine months to develop Jalapeño and that it’s the first in a multigeneration computing platform that will begin to roll out later this year and into the years ahead.
The custom chip is part of a broader industry trend among AI companies and cloud hyperscalers seeking alternatives to market leader Nvidia’s high-powered processors.
While OpenAI is one of the largest buyers of Nvidia’s chips, it has to compete with the entire AI industry to get its hands on them. Looking for secondary options, like its own processors, helps OpenAI get access to the AI computing power it needs faster.
Owning and developing more of the infrastructure necessary to run its AI models will also help OpenAI better optimize its systems for its own needs.
Like OpenAI, Amazon, Google, and Microsoft use or are working on their own custom AI processors. Meta similarly develops and deploys its own chips for its AI and other workloads.
Amazon and Google have also begun renting out those chips to third-party customers. And Meta has teased the idea of becoming a cloud computing provider, potentially increasing competition with Nvidia.
In addition to battling its own customers, Nvidia rival AMD is also working to steal away market share with its own AI data center offerings. Companies like Qualcomm and Cerebras are looking to make headway in the AI chip space.
