OpenAI and Broadcom unveil LLM-optimized inference chip

OpenAI and Broadcom have introduced Jalapeño, a custom intelligence processor optimized for LLM inference to enhance AI infrastructure performance, efficiency, and scale. This 10-gigawatt chip partnership aims to reduce costs and reshape the AI landscape through improved innovation and competition. Following the reveal of this custom hardware designed to boost AI systems, Broadcom shares rose by 3%, reflecting the industry's reaction to the new LLM-optimized hardware development.