NVIDIA LPU architecture achieves 10k+ OTSU speeds on Gemma 4
AccBalanced · x · 2026-08-25
NVIDIA's LPX team published a technical blog post detailing their latest inference architecture performance:
- Gemma 4 Performance: Reached 10,996 OTSU speeds (high AL case) and a median of 3,431 OTSU for full 100K input context window workloads.
- 2T-scale Models: Combining Vera Rubin and LPU drafts to achieve 1000 tok/s interactivity with best-in-class tok/watt efficiency.
These figures highlight NVIDIA's progress in LLM inference acceleration.
More from Infra
- Jensen Huang joins BG2 podcast: does AI revenue justify capex, '27 bottlenecks — elonmusk · 2026-10-16
- Mistral's on-device infra rack revealed: DGX Spark, Mac Studio and more — gregmushen · 2026-10-16
- NVIDIA and Sweden's Wallenberg launch Sferical One, a shared AI supercomputer — nvidia · 2026-10-16
- Next infra frontier: self-evolving systems that adapt and self-verify — _shreya_s · 2026-10-16
- Xiaomi's MiMo RL dashboard reveals $3.47M spent on just 30 training steps — bycloud · 2026-10-16
- Cerebras pushes disaggregated inference: heterogeneous prefill pools feeding wafer-scale decode — ai · 2026-10-16