Muxi launches XiJing S600 AI super node for big model training

At the 2026 World AI Conference, Muxi officially launched its new AI super node product, the XiJing S600. Designed for large model training, inference, and AI computing centers, it aims to provide an efficient domestic computing base. The XiJing S600 supports 64 GPUs in a single cabinet, uses a full-interconnect architecture to reduce data latency, and supports multiple parallel strategies. The cluster can scale to thousands of GPUs, with integrated cabinet delivery and zero-cable design to shorten deployment time and lower maintenance costs.