Charts: AMD MI455X tops per-chip inference, TPU7x beats GB300 on tokens per dollar
AccBalanced · x · 2026-11-16
@intelfabs cites two Nov 2026 charts (J.P. Morgan/Epoch AI; SemiAnalysis inference model): AMD MI455X tops modeled per-chip inference performance (3.0 GB300-equivalents), ahead of Nvidia VR200 (2.4), GB300, TPU v8i and others; and Google TPU7x @FP8 beats GB300 on token output per dollar across most speeds for Qwen 3.5 397B. The poster argues the market is diversifying, not replacing Nvidia, despite years of custom-ASIC predictions.
More from Infra
- DeepSeek Flash on Mac Studio M5 Ultra (512GB): 54 t/s generation with 128K context — vhthc · 2026-11-17
- Peter Diamandis: The Country Building Gigawatts Fastest Will Set the Price of Intelligence — PeterDiamandis · 2026-11-17
- Gemini Managed Agents Adds Self-Hosted Sandboxes: Google Runs the Loop, You Run the Tools — _philschmid · 2026-11-17
- RWKV-8 RPU preview hits 12,500+ tokens/s real-time inference — kastnerkyle · 2026-11-17
- Your AI budget is too generous and too restrictive at once: routing saves, flat caps kill rewrites — ruslansv · 2026-11-16
- Edge inference with llm-d: prefill-decode disaggregation on Qualcomm Cloud AI 100 Ultra — HowDevelop · 2026-11-16