OpenAI's chip engineering: 9 months to tapeout, AI beats humans
kleffew94 · x · 2026-08-28
OpenAI revealed details about its inference chip Jalapeño at Hot Chips. Highlights include a 9-month timeline from RTL to tapeout using XLS. An AI optimization loop autonomously improved a kernel from 0.31% to 88.94% efficiency in under two days, producing code up to 1.8x faster than hand-tuned versions and even optimizing physical hardware circuits.
Related event: OpenAI Unveils Jalapeño, Its First Custom Inference ASIC, at Hot Chips(7 posts)→
More from Infra
- How Microsoft optimized Windows for Nvidia's Arm-based RTX Spark AI laptops — tomwarren · 2026-10-15
- How Microsoft optimized Windows for Nvidia's Arm-based RTX Spark laptops — The Verge AI · 2026-10-15
- Global smartphone shipments plunge 22% in worst quarterly decline on record as AI demand fuels memory costs — Polymarket · 2026-10-15
- SK Hynix drops 11% again, now below 3x FY2027 earnings as AI memory selloff deepens — JOBhakdi · 2026-10-15
- ASML Sells Zero Lithography Machines into Europe as US Hosts 75%+ of AI Compute — NinaDSchick · 2026-10-15
- Nvidia quietly reclassified SPCX as 'Hyperscale', signaling deeper neocloud reliance — RihardJarc · 2026-10-15