OpenAI fires up “Jalapeño,” its first homegrown AI chip

OpenAI has taken a major step beyond software development by building its first custom AI chip in collaboration with Broadcom. The chip is designed specifically for AI inference tasks, meaning it helps run trained AI models more efficiently and at lower cost. This move reflects a broader industry shift where leading AI companies are trying to control not only their software systems but also the hardware that powers them. By reducing dependence on Nvidia, OpenAI aims to improve performance, reduce infrastructure costs, and increase control over its AI ecosystem as competition in the sector intensifies.