OpenAI's models wrote the code that made its new chip fast.
OpenAI's hardware chief told IEEE Spectrum how the company's own models helped build the Jalapeño chip. One benchmark went from 0.3% of the chip's peak speed to 89% in about 40 hours. OpenAI says humans still lead the work, and Broadcom handled the physical layout.
Why it mattersChips designed faster mean cheaper answers in the apps you already use. It also shows a team under 100 people attempting work that once needed hundreds.
▶ Related video: OpenAI's AI Designed Its Own Chip in 9 Months — And It Beat Nvidia!
Read in the app — free, in 9 languages
Related stories
DoorDash says AI agents have placed only a few hundred of its orders.2026-10-30A fake Claude Code update put a backdoor on up to 5,000 computers.
2026-10-30Canada is buying a stake in Cohere, its biggest homegrown AI company.
2026-10-30A worker's personal Claude agent sat on a Pentagon network for months.
2026-10-30GitHub's Copilot now picks the AI model for each coding task itself.
2026-10-29