OpenAI Unveils 'Jalapeño' Custom AI Chip for LLM Inference
What happened
OpenAI, in collaboration with Broadcom, has unveiled 'Jalapeño,' its first custom ASIC chip designed specifically for efficient LLM inference tasks. Developed in just nine months, this chip aims to reduce OpenAI's dependence on Nvidia and enhance efficiency for running models like ChatGPT and Codex.
Why it matters
Directors of AI/ML evaluating compute infrastructure should investigate custom ASIC solutions for inference workloads, as OpenAI's rapid development of Jalapeño signals a critical shift towards specialized hardware that can deliver "substantially superior" performance-per-watt.
Topics
- OpenAI
- Custom AI Chips
- ASIC Design
- AI Inference
Articles in this trend
- OpenAI's spicy new custom AI chip — The Rundown AI
- OpenAI unveils its first custom chip, built by Broadcom — AI News & Artificial Intelligence | TechCrunch
- OpenAI and Broadcom unveil "Jalapeño," a custom chip built for LLM inference — The Decoder
- OpenAI and Broadcom announce chip designed for LLM inference at scale — AI - Ars Technica
- OpenAI, Broadcom Develop AI Chip Called 'Jalapeno' — Bloomberg Technology
- OpenAI, Broadcom debut custom Jalapeño chip for AI inference — AI – SiliconANGLE
- OpenAI unveils first custom inference chip named Jalapeño — Dataconomy
- OpenAI’s Jalapeño Will Be Spicy, But the Real Sizzle Is Its Chip Design AI — Big Data & AI News - EE Times
- OpenAI’s Jalapeño chip is Big Tech’s spiciest move away from Nvidia — AI News & Artificial Intelligence | TechCrunch
- Why everyone from OpenAI to SpaceX is building their own chips (and turning up the heat on Nvidia) — AI News & Artificial Intelligence | TechCrunch
- OpenAI, Broadcom unveil custom AI chip; Qualcomm to buy Modular for $3.9 bn - Business Standard — artifical intelligence via Google News
- OpenAI pulls ahead on custom chips — Semafor