Huawei unveils Atlas 960E SuperPoD for large-scale AI computing

Chinese technology giant Huawei has unveiled a new computing system designed to support large-scale artificial intelligence model training, deployment and other high-performance computing workloads.

Called the Atlas 960E SuperPoD, the system was unveiled by Huawei Deputy Chairman and Rotating Chairman David Wang at Huawei Connect 2026 in Shanghai.

The conference brought together industry leaders, customers and technology experts to discuss AI, digital connectivity and collaboration across the technology sector.

Huawei said the Atlas 960E SuperPoD is built to handle demanding AI workloads while focusing on computing performance, energy efficiency and long-term operational reliability.

The system is designed to address the growing infrastructure requirements of large-scale AI applications.

Huawei also showcased the Atlas 950 SuperPoD, based on the first generation of its Piriam computing architecture. The company said its SuperPoD-based superclusters can link large numbers of processors to operate as a unified computing system, with the capability to coordinate up to one million processors.

The company also introduced the Taishan 950 SuperPoD for general-purpose computing and OceanStor M900, a storage solution designed to manage large volumes of AI-related data.

In another major announcement, Huawei said it would open-source its Compute Atlas network platform, allowing developers, technology partners and the wider community to contribute to its development and AI-related innovations.

Wang said stronger computing infrastructure and an open technology ecosystem are essential to sustaining the pace of AI innovation, adding that Huawei will work with partners to expand AI capabilities and create new opportunities for innovation.