tech
OpenAI Unveils Custom 'Jalapeño' AI Chip Developed With Broadcom
OpenAI, in partnership with Broadcom, has announced its first custom AI processor, codenamed 'Jalapeño.' The chip is an ASIC designed specifically for large language model inference, aiming to provide better performance per watt and reduce OpenAI's reliance on single suppliers like Nvidia. Initial deployments are expected by the end of 2026.
TL;DR
- OpenAI and Broadcom have unveiled Jalapeño, OpenAI's first custom 'Intelligence Processor' for AI.
- The chip is designed for LLM inference and claims superior performance per watt compared to current technologies.
- Jalapeño is an ASIC tuned for large language models, aiming to reduce dependence on Nvidia GPUs.
- OpenAI plans to use Jalapeño for some customer queries later this year.
- Industry observers view this as a strategic move by Big Tech to control more of the AI hardware stack and mitigate single-supplier risk.
- Some in the open-source community emphasize local and open models as an alternative.