OpenAI and Broadcom unveil custom Jalapeño chip for AI inference at scale

technology artificial intelligence hardware semiconductors

OpenAI Group PBC has revealed its first home-grown artificial intelligence chip, called Jalapeño. Developed in partnership with Broadcom Inc., the processor is designed to power the company's large language models. The reveal comes eight months after the two companies first announced a custom chip deal.

The chip was built specifically for the unique needs of OpenAI's inference systems, rather than training. By tailoring hardware to better run its AI products, the company aims to improve efficiency, lower costs, and advance its strategy to build a full stack behind its models and products.

This move is the clearest sign yet that the company most dependent on Nvidia wants a way out. The development occurs as the silicon race heats up amid the struggle to keep up with demand. Broadcom Inc. brings experience to the collaboration, having previously helped Google LLC develop its TPU line of artificial intelligence accelerators.

OpenAI and Broadcom announce chip designed for LLM inference at scale

arstechnica.com

OpenAI, Broadcom debut custom Jalapeño chip for AI inference

siliconangle.com

OpenAI, Broadcom Develop Custom Chip for AI Inference

wsj.com

OpenAI unveils its first custom chip, built by Broadcom

techcrunch.com

OpenAI built its own AI chip. The target is Nvidia.

thenextweb.com

OpenAI unveils first chip as part of Broadcom deal in effort to 'build the full stack'

cnbc.com

OpenAI, Broadcom Unveil Chip to Run Models Faster, Cheaper

bloomberg.com