SemiAnalysis 在 SemiAnalysis Weekly 节目中讨论 OpenAI 的 AI 设计芯片 Jalapeno,称其在每兆瓦 token 数上优于 Rubin,九个月内完成流片,还能运行 Doom。节目对比了 Jalapeno 与 Rubin、GB300 在 InferenceX 上的表现,并讨论为何每兆瓦 token 数才是关键指标、三星 HBM4 的回归、AI 编写的 RTL 与 kernel、CUDA 护城河是否正在松动,以及 Anthropic 的 ASIC 计划。节目观看链接为 https://t.co/C1p4oguqlt。
OpenAI's secret weapon just dropped. Jalapeno, the AI-designed chip that beats Rubin on tokens per megawatt, taped out in nine months, and runs Doom.
Myron and Bryan join @JordanNanos on SemiAnalysis Weekly to break it all down:
- Jalapeno vs Rubin and GB300 on InferenceX
- Why tokens per megawatt is the metric that matters
- Samsung's HBM4 comeback
- AI-written RTL and kernels
- Is the CUDA moat draining?
- Anthropic's ASIC is next
Watch Now:
https://t.co/C1p4oguqlt
来源:@SemiAnalysis_ · x.com