Tech
NeuraFeed25 Jun 2026OpenAI and Broadcom Unveil Jalapeño, a Custom LLM Inference Chip
OpenAI and Broadcom have officially unveiled Jalapeño, OpenAI's first custom-designed Intelligence Processor for Large Language Model (LLM) inference. This collaboration marks a significant step in OpenAI's strategy to build a full-stack AI platform, aiming to make advanced AI faster, more reliable, and more accessible. The chip, developed in just nine months, is designed to deliver substantially better performance per watt than current state-of-the-art solutions and will be deployed at gigawatt scale with data center partners.
By NeuraFeed