Supermicro unveiled its upcoming NVIDIA Vera Rubin-powered systems, including the NVL72 SuperCluster, HGX Rubin NVL8, and Vera CPU systems, designed to meet surging demand for AI inference workloads. Leveraging Supermicro's Data Center Building Block Solutions (DCBBS) liquid-cooling technology, these systems aim to accelerate deployment of AI factories. The NVL72 delivers up to 3.6 Exaflops of inference, 75 TB of fast memory, and 1.6 PB/s of HBM4 bandwidth, targeting 10x throughput per watt.

The 2U HGX Rubin NVL8 offers flexible CPU options alongside eight GPUs per rack. The Vera CPU system provides dense, energy-efficient AI compute for agentic workloads. Supermicro's DCBBS approach provides validated, pre-engineered rack solutions, reducing integration risk and lowering TCO for AI infrastructure deployments of any scale.