At the Huawei Connect event, Xu Zhijun, Deputy Chairman and Rotating CEO of Huawei, emphasized that super-nodes have emerged as the new standard in AI infrastructure development. He pointed out that large-scale super-nodes present formidable obstacles for interconnection technologies. Drawing on its robust connectivity technology expertise, Huawei has successfully surmounted this hurdle through systematic innovation. The company has introduced the super-node-centric interconnection protocol named 'Lingqu' (UnifiedBus), a solution designed to streamline and enhance super-node connectivity.

Furthermore, leveraging its cutting-edge super-node technology, Huawei has unveiled two powerful super-node clusters: the 'Atlas 950 SuperCluster' and the 'Atlas 960 SuperCluster'. These clusters boast impressive computing capabilities, with the Atlas 950 SuperCluster exceeding 500,000 cards and the Atlas 960 SuperCluster reaching an astounding one million cards, marking a significant leap in AI infrastructure performance.