ASUS today announced the rollout of XA GB721-E2, built on NVIDIA GB300 NVL72 rack-scale system. Designed for large-scale model training, high-throughput inference, and advanced AI/HPC workloads, the system combines breakthrough performance, sustainable liquid cooling, and serviceability at rack scale to help enterprises and research institutions accelerate innovation.

Rack-scale performance with NVIDIA Grace Blackwell Ultra GPU
NVIDIA GB300 NVL72 systems integrate 36 NVIDIA Grace CPUs and 72 NVIDIA Blackwell Ultra GPUs in a single NVIDIA NVLink domain, delivering ultra-low-latency, high-bandwidth GPU-to-GPU communication for trillion-parameter workloads. Together with NVIDIA Quantum-X800 InfiniBand platform or NVIDIA Spectrum-X Ethernet platform and NVIDIA ConnectX-8 SuperNIC, the system is engineered for high-throughput inference and cluster-scale expansion, fully support the recent need of enterprise AI factory innovation.