Pyyan / Compare / Trainium 3 vs B300 Blackwell Ultra vs B200 vs MI400 vs H100
Trainium 3 vs B300 Blackwell Ultra vs B200 vs MI400 vs H100
5 of 5GPU & AI Chips · verified 13 Aug 2026
| Specification | Trainium 3 | B300 Blackwell Ultra | B200 | MI400 | H100 |
|---|---|---|---|---|---|
| Summary | 30–40% better price-performance. | 15 PFLOPS dense FP4. | The current volume Blackwell part. | Double NVIDIA's memory capacity. | The workhorse of the previous generation. |
| Memory | 144GB HBM3e | 288GB HBM3e | 192GB HBM3e | 432GB HBM4 | 80GB HBM3 |
| Bandwidth | — | 8 TB/s | — | 19.6 TB/s | — |
| Price | — | ~$40–50K (verify) | $30–50K | — | $25–40K |
| Vendor | AWS | NVIDIA | NVIDIA | AMD | NVIDIA |
| Category | GPU & AI Chips | GPU & AI Chips | GPU & AI Chips | GPU & AI Chips | GPU & AI Chips |
| Official | AWS ↗ | NVIDIA ↗ | NVIDIA ↗ | AMD ↗ | NVIDIA ↗ |
Highlighted rows are where these differ.
B300 Blackwell Ultra
- 8 TB/s memory bandwidth
- Per-GPU pricing widely misquoted — one source cites rack pricing
Best for frontier training.
MI400
- 19.6 TB/s bandwidth
- 40 PFLOPS FP4
- $7.2B in commitments
Best for large-model inference.