Huawei''s CloudMatrix 384: A Bold AI Architecture Challenging Nvidia''s
Unlike Nvidia''s GPU-centric approach, Huawei employs an all-to-all topology, connecting 384 Ascend 910C chips through advanced optical interconnects rather than traditional copper links.
Inside Huawei CloudMatrix 384: 2xThroughput of NVIDIA GB200
The Huawei CloudMatrix 384 is a bold, strategic counterpunch to NVIDIA, proving that domestic engineering, scale, and optical networking can offset per-chip performance disadvantages.
Huawei CloudMatrix 384 Super-node: 6912×400G SiPh LPO Optics
CloudMatrix 384 super-node uses 6912 x 400G OSFP silicon photonic (SiPh) Linear Drive Pluggable Optics (LPO) optical modules and 3168 fibers, connecting 384 Ascend 910C
Huawei''s AI CloudMatrix 384 is big, expensive and power hungry but it
The CloudMatrix 384 doesn''t only outperform the NVL72 in terms of compute, it also offers 3.6x aggregate memory capacity and 2.1x more memory bandwidth.
Huawei''s SuperNode CloudMatrix384: Networking
Huawei''s 910C expansion relies on a hybrid optical + electrical interconnect, whereas NVIDIA primarily uses electrical interconnects. As
Huawei CloudMatrix 384: China''s AI Powerhouse vs Nvidia
Discover how Huawei''s CloudMatrix 384 AI supercomputer outpaces Nvidia''s GB200 NVL72 with 384 Ascend chips, massive bandwidth, and optical interconnect.
Huawei AI CloudMatrix 384 – China''s Answer to Nvidia GB200 NVL72
Meet China''s newest and most powerful Chinese domestic solution, the CloudMatrix 384 built using the Ascend 910C. This solution competes directly with the GB200 NVL72, and in some
Huawei''s CloudMatrix 384: A Bold AI Architecture
Unlike Nvidia''s GPU-centric approach, Huawei employs an all-to-all topology, connecting 384 Ascend 910C chips through advanced optical
Huawei''s CloudMatrix 384 Roars Past Nvidia: China''s AI Giant
A fully optical interconnect backbone, replacing traditional copper links and enabling ultra-low latency communication between nodes. This isn''t just a larger machine; it''s a more capable,
Optical Modules in Intelligent Computing Scenarios
In the AI era, Huawei provides a full range of GE to 800GE optical modules, featuring three major capabilities: Spanning (ultra-long transmission), Stable (ultra-high reliability), and Secure (ultra-solid
Play with optical interconnection, Huawei Ascend 384 super node
The optical interconnection of Ascend 384 uses the Nebula optical module specially developed by Huawei HiSilicon for the optical interconnection scenarios of intelligent computing centers.
Huawei''s new AI CloudMatrix cluster beats Nvidia''s GB200 by brute
At glance Huawei''s CloudMatrix 384 is a rack-scale AI system composed of 384 Ascend 910C processors arranged in a fully optical, all-to-all mesh network.
Micro-Modular & Edge DC
Prefabricated micro-modular data centers and edge pods, scalable from 5 to 50 racks, ready for 5G and edge AI workloads.
Immersion & Liquid Cooling
Single-phase immersion cooling tanks and direct-to-chip liquid cooling switches, achieving PUE below 1.1.
AI Servers & Racks
GPU-accelerated AI servers, high-density server racks, and network cabinets optimized for AI/ML workloads.
DCIM/EMS & Cable Bridge
Real-time data center infrastructure management, plus overhead cable trays and fiber bridges for structured cabling.