Developing5 outlets · over 9 hours
Why OpenAI's Jalapeño Might Sicken Nvidia
OpenAI's Jalapeño inference ASIC, co-developed with Broadcom, delivered 1.5x to 1.9x more throughput per kilowatt and 1.7x to 3.6x lower latency than Nvidia's GB200 and GB300 on SemiAnalysis's InferenceX suite. Deployment in OpenAI data centers begins later this year.
The story so far2 updates, latest first.
Tracked by ByteBrief