Skip to content

OpenAI Launches Custom Chip With Broadcom To Reduce Nvidia Dependence

The chip will be manufactured by Broadcom and used primarily for AI inference, the computing process that powers responses in ChatGPT and other applications. OpenAI President Greg Brockman sai

Pic via @OpenAI

OpenAI and Broadcom on Wednesday unveiled Jalapeño, the ChatGPT maker’s first custom artificial intelligence chip, marking a major step in its effort to build more of the technology stack behind its AI products.

📱
According to a joint announcement, the chip will be manufactured by Broadcom and used primarily for AI inference, the computing process that powers responses in ChatGPT and other applications.

OpenAI President Greg Brockman said the project is aimed at improving efficiency, lowering costs, and expanding access to advanced AI capabilities.

The report said OpenAI has long relied on Nvidia’s graphics processing units but is seeking alternative hardware as demand for AI services continues to surge. The company has also signed agreements with Amazon Web Services, AMD, and Cerebras to diversify its infrastructure.

Jalapeño is an application-specific integrated circuit designed for targeted AI workloads. OpenAI said the chip was developed in nine months, with initial deployment expected by the end of 2026. Broadcom shares rose following the announcement.

Related Tweet:

Also Read:

Meta Bets On Affordable AI Wearables With New Smart Glasses
The new Meta Glasses start at $299, undercutting the company’s Ray-Ban-branded models, which begin at $379

Comments

Latest