OpenAI's In-House Chip Reportedly Beats Nvidia Blackwell on Efficiency
Leaks and SemiAnalysis testing indicate OpenAI's custom ASIC Jalapeño, with a 700W TDP, roughly doubles Nvidia Blackwell's throughput per watt in most no-finetuning inference scenarios, going from RTL to tape-out in just 9 months.
2026-08-26 ~ 2026-08-26 · 3 related posts
- Episode 1: OpenAI Unveils First In-House Inference Chip Jalapeño, Claiming Gains Over Nvidia Blackwell(2026-08-25, 48 posts)
- Episode 2: OpenAI to Deploy Its Jalapeño Chip by Year-End with Next Generations in Development(2026-08-26, 2 posts)
- Episode 3: OpenAI's Jalapeño Chip Play: Winning Either Way(2026-08-26, 2 posts)
- Episode 4: OpenAI's In-House Chip Reportedly Beats Nvidia Blackwell on Efficiency(2026-08-26, 3 posts)
- OpenAI's Jalapeño Chip Reportedly Beats Nvidia Blackwell on Perf/Watt — petrusenko_max · 2026-08-26
- OpenAI's 'Jalapeno' chip reportedly beats Nvidia Blackwell in efficiency and latency benchmarks — 量子位 · 2026-08-26
- OpenAI's custom chip 'Jalapeño' reportedly beats Nvidia Blackwell in efficiency — yacineMTB · 2026-08-26