Pyyan / Compare / Groq LPU vs B300 Blackwell Ultra vs B200 vs MI400 vs Ascend 950DT
Groq LPU vs B300 Blackwell Ultra vs B200 vs MI400 vs Ascend 950DT
5 of 5GPU & AI Chips · verified 24 Sept 2026
| Specification | Groq LPU | B300 Blackwell Ultra | B200 | MI400 | Ascend 950DT |
|---|---|---|---|---|---|
| Summary | Inference only, ~18× GPU speed. | 15 PFLOPS dense FP4. | The current volume Blackwell part. | Double NVIDIA's memory capacity. | 144GB on Huawei's own HBM, and the part China is building gigawatts around. |
| Memory | — | 288GB HBM3e | 192GB HBM3e | 432GB HBM4 | 144GB HiZQ 2.0 |
| Bandwidth | — | 8 TB/s | — | 19.6 TB/s | 4 TB/s |
| Price | — | ~$40–50K (verify) | $30–50K | — | Above ¥250,000 (about $37K), reported |
| Vendor | Groq | NVIDIA | NVIDIA | AMD | Huawei |
| Category | GPU & AI Chips | GPU & AI Chips | GPU & AI Chips | GPU & AI Chips | GPU & AI Chips |
| Official | Groq ↗ | NVIDIA ↗ | NVIDIA ↗ | AMD ↗ | Huawei ↗ |
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.
Ascend 950DT
- 4 TB/s memory bandwidth and a 2 TB/s interconnect, 2.5x the Ascend 910C
- HiZQ 2.0 is Huawei's own memory standard rather than a licensed HBM3 or HBM3e part
- Indicated price rose 20% to 50% in two months to September 2026, on a memory shortage
- Successor Ascend 960DT pulled forward to Q1 2027, Huawei said on 17 September
Best for serving models where Nvidia cannot be bought.