OpenAI unveils first custom AI inference chip, Jalapeño, with Broadcom — and its development was sped-up with OpenAI's own models

OpenAI and Broadcom have teamed up to introduce the Jalapeño chip, a specialized AI accelerator for handling large language model inference tasks. This marks a shift from the more generic GPUs used by companies like Nvidia and AMD. The development of Jalapeño was accelerated by leveraging OpenAI's own models, indicating a strategic move to enhance efficiency in processing tasks for products like ChatGPT. This innovation could potentially offer more powerful and cost-effective solutions for AI-driven applications, signaling a new frontier in AI hardware collaboration.