OpenAI finally has its own artificial-intelligence chip, joining an increasingly crowded field of custom chip makers.
On Wednesday, the ChatGPT maker unveiled an AI processor called Jalapeño -- yes, like the pepper -- which incorporates silicon from Broadcom and rack technology from Celestica. OpenAI designed the chip specifically for AI inference, which refers to using trained large-language models to execute tasks and solve problems.
OpenAI and Broadcom first announced a partnership to build AI accelerators last October.
Broadcom stock rose 0.8% in early trading Wednesday, while Celestica surged 3.2%. The broader iShares Semiconductor exchange-traded fund was down 0.4%.
OpenAI has been a prolific user of advanced chips. CEO Sam Altman declared a year ago that the company would have over a million graphics processing units, or GPUs, online by the end of 2025. It now joins peers like Alphabet, Amazon, and Meta Platforms in producing custom chips, which could reduce its reliance on GPU leader Nvidia and lower costs.
"By designing more of the stack ourselves, we can serve more intelligence with greater efficiency and keep pushing advanced AI toward broader access," said OpenAI President Greg Brockman.
Jalapeño will be part of a broader AI platform set for initial deployment by the end of the year. It will eventually be used "at gigawatt scale" with OpenAI's data-center partners, the company said.
Write to Nate Wolf at nate.wolf@barrons.com
This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
June 24, 2026 10:44 ET (14:44 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.