OpenAI Unveils First Custom AI Chip, Jalapeño
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- OpenAI and Broadcom unveiled Jalapeño, their first custom LLM-optimized inference chip, completing the design-to-manufacturing tape-out cycle in nine months.
- Jalapeño's development was sped up using OpenAI's own AI models, with early testing showing performance per watt 'substantially better than current state-of-the-art' (per OpenAI).
- Coverage frames the chip as OpenAI's bid to cut dependence on Nvidia and pursue an 'Apple-like move to build the full stack' (TechRadar), with targets of 50% cheaper inference and end-of-2026 deployment (Tech Times, Fortune India).
Why it matters: A first custom inference chip, built in nine months with AI-assisted design and targeting 50% cheaper inference, gives OpenAI a direct lever to reduce Nvidia dependence and control its own compute economics by the end of 2026.
Ask SkimNews
