Huawei Ascend 910C
Domestic training and inference silicon with the CANN/MindSpore stack and CloudMatrix racks.
Full Huawei Ascend 910C record →Head to head
Huawei Ascend 910C (Huawei) and NVIDIA Blackwell (NVIDIA) both sit in Chips & Compute. NVIDIA Blackwell shipped more tracked changes in the last 30 days (8 vs 4). Every value below comes from the latest crawl of the vendors' own pages.
| Attribute | Huawei Ascend 910C Huawei | NVIDIA Blackwell NVIDIA |
|---|---|---|
| Segment | Chips & Compute | Chips & Compute |
| Version | Ascend 910C | Blackwell |
| Entry cost | Pricing is not publicly available (enterprise hardware) | Pricing is not publicly listed (enterprise hardware architecture/system) |
| Pricing tiers | — | — |
| Model stack | Ascend 910C · CANN / MindSpore | Blackwell · Blackwell B200 · Grace CPU |
| Context window | — | — |
| Public API | ||
| Routes models | ||
| Changes / 30d | 4 | 8 |
| Origin | China | Global |
| Capabilities | dual-die 910C, CloudMatrix 384 racks, CANN toolkit, MindSpore, CANN 异构计算架构, MindSpore AI框架, MindStudio 全流程工具链, NPU | 72-GPU NVLink domain, FP4 inference, HBM3E 192GB, CUDA ecosystem, Second-generation Transformer Engine, Fifth-generation NVLink, FP4 AI Support, Tensor Cores with microscaling formats |
Domestic training and inference silicon with the CANN/MindSpore stack and CloudMatrix racks.
Full Huawei Ascend 910C record →NVL72 rack-scale coherent GPU domains, the default substrate for frontier training and inference.
Full NVIDIA Blackwell record →