OpenAI Develops Custom AI Chips with Broadcom

OpenAI is advancing its AI capabilities by developing custom chips in collaboration with Broadcom and Taiwan Semiconductor Manufacturing Company (TSMC). This strategic move aims to secure a reliable supply chain and manage costs amid increasing computational demands and supply shortages from Nvidia. OpenAI's new chip design will focus on AI inference, complementing existing training chips from Nvidia and AMD. The company has assembled a team of about 20 engineers, including those experienced with Google's TPUs, to develop its first custom AI chip by 2026. By diversifying its chip supply and focusing on in-house design, OpenAI aligns with strategies of tech giants like Amazon, Meta, Google, and Microsoft. This initiative reflects OpenAI's commitment to maintaining its competitive edge in the AI sector.