Mars Finance news. April 7, 2026. Recently, the first “Zhenwu” 10,000-card intelligent computing cluster for the Guangdong-Hong Kong-Macao Greater Bay Area, jointly built by China Telecom and Alibaba Cloud, was officially launched at the Shaoguan data center. As the first implementation project of the national “ultra-large-scale intelligent computing cluster” new infrastructure initiative in the Greater Bay Area, the cluster achieves end-to-end independent R&D across the entire chain—from chips, to the cloud platform, to model applications—marking that China’s advanced computing power is moving from breakthroughs in high-end performance toward large-scale industrial deployments. In terms of technology, the cluster achieves multiple key breakthroughs. Through technological innovations such as high-performance networking between cards via RoCE and dual-plane multi-track communication, the cluster’s end-to-end network latency is as low as 4 microseconds, and network peak utilization exceeds 95%. It can efficiently meet the communication needs of super-large traffic such as AllReduce and AlltoAll during large model training, and stably support the pre-training and inference tasks of large models with hundreds of billions of parameters. The Zhenwu chip turnkey system provides 1.5T of large display memory, and the interconnect bandwidth between cards exceeds 700GB/s, providing solid support for the domestic computing power foundation. (Wide-angle observation)
粤港澳大湾区初の「真武」万卡智算クラスターが稼働開始
Mars Finance news. April 7, 2026. Recently, the first “Zhenwu” 10,000-card intelligent computing cluster for the Guangdong-Hong Kong-Macao Greater Bay Area, jointly built by China Telecom and Alibaba Cloud, was officially launched at the Shaoguan data center. As the first implementation project of the national “ultra-large-scale intelligent computing cluster” new infrastructure initiative in the Greater Bay Area, the cluster achieves end-to-end independent R&D across the entire chain—from chips, to the cloud platform, to model applications—marking that China’s advanced computing power is moving from breakthroughs in high-end performance toward large-scale industrial deployments. In terms of technology, the cluster achieves multiple key breakthroughs. Through technological innovations such as high-performance networking between cards via RoCE and dual-plane multi-track communication, the cluster’s end-to-end network latency is as low as 4 microseconds, and network peak utilization exceeds 95%. It can efficiently meet the communication needs of super-large traffic such as AllReduce and AlltoAll during large model training, and stably support the pre-training and inference tasks of large models with hundreds of billions of parameters. The Zhenwu chip turnkey system provides 1.5T of large display memory, and the interconnect bandwidth between cards exceeds 700GB/s, providing solid support for the domestic computing power foundation. (Wide-angle observation)