Back

Google's TPUv7 Ironwood delivers up to 50% better inference economics than Nvidia chips

Verifying
  • Google's TPUv7 Ironwood outperformed Nvidia's B200 and B300 on inference cost efficiency.
  • Ironwood delivered 50.4% more tokens per dollar than B200 and 96.0% more than B300.
  • SemiAnalysis tested Qwen3.5 397B FP8 in aggregate serving mode.
The information on this website is generated using AI and we cannot guarantee its accuracy. Please use it as reference information only.