Tomorrow
← Chips & Compute

Huawei Ascend 910C

Huawei · #7 most active of 18 in Chips & Compute

Compare →

Domestic training and inference silicon with the CANN/MindSpore stack and CloudMatrix racks.

Not yet checkedModel stack: UnverifiableVersion: UnverifiableHow this works →

Current version

Ascend 910C

Entry cost

Pricing is not publicly available (enterprise hardware)

Changes / 30d

4

The short answers · verified September 17, 2026

What is the latest version of Huawei Ascend 910C?
The current shipped version of Huawei Ascend 910C is Ascend 910C, as of September 17, 2026.
How much does Huawei Ascend 910C cost?
Huawei Ascend 910C has no published flat entry price. Pricing is not publicly available (enterprise hardware)
What AI model does Huawei Ascend 910C use?
Huawei Ascend 910C runs primarily on Ascend 910C.

Source: Huawei — www.hiascend.com · Reusable under CC BY 4.0 — cite Tomorrow

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
  • AI Acceleration
  • CANN
  • MindStudio
  • MindIE
  • MindSpeed
  • NPU processor
  • AI Computing Platform
  • Heterogeneous Computing Architecture (CANN)
  • Deep Learning Model Training and Inference
  • AI Hardware
  • NPU Processor
  • Heterogeneous Computing
  • High Performance Computing
  • High-performance computing
  • AI Training
  • AI Inference
  • NPU Accelerator
  • AI Processing
  • NPU Hardware
  • AI training
  • AI inference
  • Heterogeneous computing
  • AI Compute
  • Hardware Accelerator
  • AI Computing
  • NPU Acceleration
  • Accelerated Computing
  • AI hardware acceleration
  • Model training & inference support
  • NPU hardware
  • AI Chip
  • Hardware Acceleration
  • AI Processor
  • CANN Architecture
  • CANN Heterogeneous Compute Architecture
  • MindSpore support
  • PyTorch support
  • TensorFlow support
  • AI hardware
  • AI Training & Inference

Pricing

Pricing is not publicly available (enterprise hardware)

Entry price over time

Not enough pricing history yet — we start charting from the second observation.

Geek mode

Models under the hood

  • Ascend 910C

    Huawei

    disclosed

  • CANN / MindSpore

    Huawei

    disclosed

Context window

—

Public API

no

Multi-model routing

no

Stack signals

  • CANN
  • MindSpore
  • PyTorch
  • TensorFlow

Shared model stack

Other tracked products running on the same foundation models — a quick read on how much of the catalog moves when one of these models changes.

green disclosed · amber inferred · grey unattributed. Aliases are folded into one model; provider concentration counts only models with an established vendor.

Reported scale

Reference data. Each figure is whatever the source actually said — weekly users, downloads, revenue run-rate — with its own definition and date. These are not comparable between tools and are never used to rank anything.

No public usage figure on record for this product.

Also in Chips & Compute

All Chips & Compute
Tenstorrent Blackhole

Tenstorrent

15

Open RISC-V AI hardware

Version
Blackhole
Cost
from $999/unit
Model
Tensix cores
Google TPU Ironwood

Google

10

Inference-first TPU pods

Version
Ironwood (7th generation)
Cost
from $5.4/unit
Model
Gemini
Cerebras WSE-3

Cerebras

8

Wafer-scale inference and training

Version
WSE-3
Cost
Free tier / developer tier pricing available (contact/partner integrations)
Model
Codex-Spark
NVIDIA Blackwell

NVIDIA

8

GB200/B200 rack-scale AI systems

Version
Blackwell
Cost
Pricing is not publicly listed (enterprise hardware architecture/system)
Model
Blackwell
AWS Trainium2

AWS

5

Cloud-native training silicon

Version
Trainium2
Cost
Pricing not published; available via AWS EC2 instance pricing.
Model
Neuron compiler
Groq LPU

Groq

4

Deterministic low-latency inference

Version
LPU
Cost
Pricing not published. Contact sales or check account console for usage-based rates.
Model
canopylabs/orpheus-arabic-saudi

Change history

  • capability

    New capabilities: AI Training & Inference

    47 tracked48 tracked · +AI Training & Inference

    source
  • capability

    New capabilities: AI hardware

    46 tracked47 tracked · +AI hardware

    source
  • capability

    New capabilities: CANN Heterogeneous Compute Architecture, MindSpore support, PyTorch support

    42 tracked46 tracked · +CANN Heterogeneous Compute Architecture, MindSpore support, PyTorch support, TensorFlow support

    source
  • capability

    New capabilities: AI Processor, CANN Architecture

    40 tracked42 tracked · +AI Processor, CANN Architecture

    source
  • capability

    New capabilities: AI Chip, Hardware Acceleration

    38 tracked40 tracked · +AI Chip, Hardware Acceleration

    source
  • capability

    New capabilities: NPU hardware

    37 tracked38 tracked · +NPU hardware

    source
  • capability

    New capabilities: AI hardware acceleration, Model training & inference support

    35 tracked37 tracked · +AI hardware acceleration, Model training & inference support

    source
  • capability

    New capabilities: Accelerated Computing

    34 tracked35 tracked · +Accelerated Computing

    source
  • version

    Huawei Ascend 910C moved to Ascend 910C

    910CAscend 910C

    source
  • capability

    New capabilities: NPU Acceleration

    33 tracked34 tracked · +NPU Acceleration

    source
  • capability

    New capabilities: AI Computing

    32 tracked33 tracked · +AI Computing

    source
  • version

    Huawei Ascend 910C moved to 910C

    Ascend 910C910C

    source
  • version

    Huawei Ascend 910C moved to Ascend 910C

    910CAscend 910C

    source
  • capability

    New capabilities: AI Compute, Hardware Accelerator

    30 tracked32 tracked · +AI Compute, Hardware Accelerator

    source
  • version

    Huawei Ascend 910C moved to 910C

    Ascend 910C910C

    source
  • capability

    New capabilities: AI training, AI inference, Heterogeneous computing

    27 tracked30 tracked · +AI training, AI inference, Heterogeneous computing

    source
  • version

    Huawei Ascend 910C moved to Ascend 910C

    910CAscend 910C

    source
  • capability

    New capabilities: AI Processing, NPU Hardware

    25 tracked27 tracked · +AI Processing, NPU Hardware

    source
  • version

    Huawei Ascend 910C moved to 910C

    Ascend 910C910C

    source
  • capability

    New capabilities: AI Training, AI Inference, NPU Accelerator

    22 tracked25 tracked · +AI Training, AI Inference, NPU Accelerator

    source
  • capability

    New capabilities: High-performance computing

    21 tracked22 tracked · +High-performance computing

    source
  • capability

    New capabilities: Heterogeneous Computing, High Performance Computing

    19 tracked21 tracked · +Heterogeneous Computing, High Performance Computing

    source
  • version

    Huawei Ascend 910C moved to Ascend 910C

    910CAscend 910C

    source
  • capability

    New capabilities: AI Hardware, NPU Processor

    17 tracked19 tracked · +AI Hardware, NPU Processor

    source
  • capability

    New capabilities: NPU processor, AI Computing Platform, Heterogeneous Computing Architecture (CANN)

    134

    source
  • version

    Huawei Ascend 910C moved to 910C

    Ascend 910C910C

    source
  • capability

    New capabilities: NPU, AI Accelerators, Heterogeneous Computing Architecture

    3 tracked6 tracked · +NPU, AI Accelerators, Heterogeneous Computing Architecture

    source
  • capability

    New capabilities: CANN, MindStudio, MindIE

    95

    source
  • capability

    New capabilities: NPU, AI Acceleration

    72

    source
  • version

    Huawei Ascend 910C moved to 910C

    Ascend 910C910C

    source
  • capability

    New capabilities: CANN 异构计算架构, MindSpore AI框架, MindStudio 全流程工具链

    43

    source
  • capability

    New capabilities: Hardware Accelerator

    2 tracked3 tracked · +Hardware Accelerator

    source
  • capability

    CloudMatrix 384 rack systems enter wider domestic deployment

    —CloudMatrix 384

    source
  • capability

    New capabilities: AI Training, AI Inference

    3 tracked5 tracked · +AI Training, AI Inference

    source
  • capability

    New capabilities: AI Hardware, CANN Architecture

    3 tracked5 tracked · +AI Hardware, CANN Architecture

    source
  • capability

    New capabilities: AI Accelerator, NPU, Heterogeneous Computing

    4 tracked7 tracked · +AI Accelerator, NPU, Heterogeneous Computing

    source
  • capability

    New capabilities: AI Computing, Ascend Architecture, CANN Platform

    3 tracked6 tracked · +AI Computing, Ascend Architecture, CANN Platform

    source
  • capability

    New capabilities: Heterogeneous Computing Architecture, Model Training and Inference

    5 tracked7 tracked · +Heterogeneous Computing Architecture, Model Training and Inference

    source
  • capability

    New capabilities: AI Hardware, Heterogeneous Computing, CANN Architecture

    4 tracked9 tracked · +AI Hardware, Heterogeneous Computing, CANN Architecture, Deep Learning Training, Inference

    source
  • capability

    New capabilities: AI hardware, Model training, Model inference

    3 tracked7 tracked · +AI hardware, Model training, Model inference, Heterogeneous computing

    source
Subscribe to Huawei Ascend 910C changes

Huawei Ascend 910C compared

Straight head-to-head pages against the busiest products in Chips & Compute.

Huawei Ascend 910C vs Tenstorrent BlackholeHuawei Ascend 910C vs Google TPU IronwoodHuawei Ascend 910C vs Cerebras WSE-3Huawei Ascend 910C vs NVIDIA BlackwellHuawei Ascend 910C vs AWS Trainium2Huawei Ascend 910C vs Groq LPU

How to cite this page

Free to cite and reuse under CC BY 4.0. Permalink: https://tomorrow.aliensquad.ai/tools/huawei-ascend

APA
Tomorrow. (2026). Huawei Ascend 910C — version, pricing and model stack [Data set entry]. AlienSquad. Retrieved 2026-09-17, from https://tomorrow.aliensquad.ai/tools/huawei-ascend
BibTeX
@misc{tomorrow-tools-huawei-ascend,
  author       = {{Tomorrow}},
  title        = {Huawei Ascend 910C — version, pricing and model stack},
  year         = {2026},
  publisher    = {AlienSquad},
  howpublished = {\url{https://tomorrow.aliensquad.ai/tools/huawei-ascend}},
  note         = {Accessed: 2026-09-17}
}