Huawei logo vs

Huawei Ascend 910C vs NVIDIA H20 96GB

The Ascend 910C delivers 5.4x the FP16 throughput of the H20 (800 vs 148 TFLOPS dense). The Ascend 910C also carries 32 GB more memory (128 GB vs 96 GB).

Ascend 910C: Da Vinci v2 (dual-die), 2025 H20: Hopper, 2024
FP16 dense lead
5.4x
Ascend 910C: 800 vs 148 TFLOPS
Memory
128 vs 96 GB
32 GB more for the Ascend 910C
Bandwidth
3.2 TB/s vs 4.0 TB/s
H20 moves data faster
TDP
600 W vs 400 W
H20 draws 200 W less
Download comparison sheet

Share this page

Huawei Ascend 910C vs NVIDIA H20 96GB

A post with this link shows the card on the Image tab as its preview.

Printable version →
Huawei logo

Huawei Ascend 910C

Leads on FP16 Full specs →

What is the Huawei Ascend 910C?

The Ascend 910C is an NPU from Huawei on the Da Vinci v2 (dual-die) architecture, an accelerator built for AI inference. It is an OAM module, the open accelerator format used in multi-GPU servers. Ascend 910C full specifications →

Can I rent the Ascend 910C in the cloud?

Flopper is not tracking any live rental listings for the Ascend 910C yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Ascend 910C compare with other accelerators?

This page sets the Ascend 910C against the H20. The Ascend 910C is the newer generation of the two. This is a cross-vendor matchup, Huawei against NVIDIA; the Ascend 910C is an OAM module while the H20 is an SXM module. The compare tool builds the same comparison for any two parts in the catalogue. All GPU comparisons →

Where do the Ascend 910C figures come from?

Every Ascend 910C figure on this page traces to Huawei's own datasheet or specification table. How Flopper verifies specifications →

NVIDIA H20 96GB

Full specs →

What is the NVIDIA H20?

The H20 is a GPU from NVIDIA on the Hopper architecture, an accelerator for AI training, inference and high-performance computing. It is an SXM module, the mezzanine format used in dense multi-GPU servers. H20 full specifications →

Can I rent the H20 in the cloud?

Flopper is not tracking any live rental listings for the H20 yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the H20 compare with other accelerators?

This page sets the H20 against the Ascend 910C. The H20 is the older generation of the two. This is a cross-vendor matchup, NVIDIA against Huawei; the H20 is an SXM module while the Ascend 910C is an OAM module. Flopper also compares it head to head with the Zhenwu 810E, across every published precision, memory and power. H20 vs Zhenwu 810E →

Where do the H20 figures come from?

Every H20 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Specifications

Huawei logo
Ascend 910C
H20
Architecture
Da Vinci v2 (dual-die)Hopper
Launch Year
20252024
Form Factor
OAMSXM
Memory
128 GB96 GB
Memory Bandwidth
3.2 TB/s4.0 TB/s
TDP
600 W400 W
Process Node
SMIC 7nm4nm

Performance (TFLOPS)

Huawei logo
Ascend 910C
H20
No verified data available1 TFLOPS
50 TFLOPS 44 TFLOPS
No verified data available74 TFLOPS
sparse not published
800 TFLOPS 148 TFLOPS
sparse not published
800 TFLOPS 148 TFLOPS
sparse not published
No verified data available296 TFLOPS
592 TFLOPS sparse
FP6
No verified data available No verified data available
No verified data available No verified data available
1,600 TOPS 296 TOPS
sparse not published

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The Ascend 910C is the newer part: Da Vinci v2 (dual-die) (2025) against the H20's Hopper (2024).

The Ascend 910C carries 32 GB more memory (128 GB against 96 GB), so larger models fit on one GPU before you shard. Bandwidth is 3.2 TB/s against 4.0 TB/s, which sets the ceiling on decode speed.

Power is 600 W against 400 W, and at FP32 the H20 does more work per watt. Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP64, FP32, TF32, BF16, FP16, FP8, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

Is the Ascend 910C faster than the H20?

At FP16 precision the Ascend 910C reaches 800 TFLOPS dense against 148 TFLOPS for the H20. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the Ascend 910C or the H20?

Choose the Ascend 910C for training and high-throughput inference: it has 5.4x the FP16 throughput of the H20. It also carries 32 GB more memory, so larger models fit on one GPU before you shard. The Ascend 910C draws 200 W more power, so budget for cooling and power delivery before you commit. Packaging differs too: the Ascend 910C is an OAM module while the H20 is an SXM module, so check what your servers accept.

Which has more memory, the Ascend 910C or the H20?

The Ascend 910C carries 128 GB of memory versus 96 GB for the H20. Memory bandwidth is 3.2 TB/s for the Ascend 910C and 4.0 TB/s for the H20.

How much power do the Ascend 910C and the H20 draw?

The Ascend 910C is rated at 600 W TDP and the H20 at 400 W. On FP32 throughput per watt, the H20 is the more efficient part.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

© 2026 Flopper.io - Compare the GPUs Powering AI