Atlas
Huawei's line of AI training and inference processors (NPUs) and the systems built on them -- the platform brand for the silicon itself (still called Ascend in some regional markets and in the underlying chip generation names), sold standalone and in Atlas-branded servers and SuperPoD clusters, now recorded in their own separate hardware and software-stack products.
Sold to data-centre and OEM customers under negotiated commercial terms; hiascend.com, Huawei's own Ascend product portal, lists specifications and cluster configurations (Atlas 950 SuperPoD, Atlas 650E AI server) but no public price.
What it does
- Accelerator silicon
- Model training
- Model inference
How it compares
-
NVIDIA Blackwell
suppliers and licensors of hardware and software for discrete and integrated GPUs, custom chips and other accelerated computing solutions, including solutions offered for AI, such as Advanced Micro Devices, Inc., or AMD, Huawei Technologies Co. Ltd., or Huawei, and Intel Corporation, or Intel
Regulatory filing · 18 Sep 2026
Sources
- Sold within
-
Huawei's Ascend portal presents the Atlas line through specifications, partner hardware and industry solutions, and its commercial control on every product page is "Project Consulting". No price, rate card or order route appears anywhere on the English site.
- Name
-
Official documentation · 1 Sep 2026
Huawei's English-language properties now brand the whole stack Atlas, not Ascend -- hiascend.com titles itself 'Atlas Community', CANN is described as 'the core of Atlas AI's basic software and hardware platform', MindStudio 'equips Atlas AI developers', and the silicon itself is written '950DT NPU', never 'Ascend 950'. 'Ascend' survives in the domain (hiascend.com), the Chinese brand 昇腾, and some regional pages. Confirmed live 2026-09-01.
Also from Huawei
-
Atlas 650E AI Server Hardware
14U AI server powered by eight Huawei 950DT NPUs, rated at up to 12.4 PFLOPS at mxFP4, for on-premises AI training and inference in finance, government and healthcare deployments.
-
Atlas 950 SuperPoD Hardware
Rack-scale AI supercomputing cabinet built from 64 Huawei 950DT NPUs per cabinet and scalable to 1,024 NPUs over a UB Link fabric for trillion-parameter model training and inference.
-
CANN Developer tool
Huawei's heterogeneous compute architecture for its Ascend/Atlas NPUs, supplying the operator libraries, compiler and programming interfaces that bridge AI frameworks to the hardware.
-
MindIE Developer tool
Inference engine and serving framework for Atlas/Ascend hardware that deploys LLM and diffusion models behind unified APIs compatible with vLLM, OpenAI and Triton interfaces.
-
MindSpore Developer tool
Open-source AI framework originated by Huawei for building, training and deploying models with native distributed training, best optimised for Huawei's Ascend/Atlas processors.
-
MindStudio Developer tool
End-to-end development toolchain for Atlas/Ascend AI applications, covering custom operator development, model conversion and compression, accuracy debugging and performance profiling via MindStudio Insight.
Something wrong here? Send a correction — quote this product id: huawei-ascend.