Supermicro begins full manufacturing production of rack-scale systems with NVIDIA's latest HGX B200 Blackwell accelerators.

Supermicro, Inc. (NASDAQ: SMCI), a Total IT Solution Provider for AI/ML, HPC, Cloud, Storage, and 5G/Edge, has announced full production availability of its end-to-end AI data center Building Block Solutions accelerated by the NVIDIA Blackwell platform. The portfolio provides core infrastructure elements for scaling Blackwell solutions with exceptional time to deployment, including a broad range of air-cooled and liquid-cooled systems with multiple CPU options, offering superior thermal design supporting traditional air cooling, liquid-to-liquid (L2L), and liquid-to-air (L2A) cooling. The offerings include full data center management software, rack-level integration with network switching and cabling, and cluster-level L12 solution validation available as turn-key offerings with global delivery, professional support, and service.

The new liquid-cooled 4U NVIDIA HGX B200 8-GPU system features newly developed cold plates and a new 250kW coolant distribution unit (CDU) that more than doubles the cooling capacity of the previous generation while maintaining the same 4U form factor. Available in 42U, 48U, or 52U rack configurations, the new vertical coolant distribution manifolds (CDM) no longer occupy valuable rack units, enabling 8 systems comprising 64 NVIDIA Blackwell GPUs in a 42U rack, and up to 12 systems with 96 NVIDIA Blackwell GPUs in a 52U rack. Supermicro also offers an in-row CDU option for large deployments, as well as liquid-to-air cooling rack solutions that do not require facility water.

For traditional data centers, the new 10U air-cooled NVIDIA HGX B200 system features a redesigned chassis with expanded thermal headroom to accommodate eight 1000W TDP Blackwell GPUs. Up to 4 of the new 10U air-cooled systems can be installed and fully integrated in a rack, the same density as the previous generation, while providing up to 15x inference and 3x training performance. The new SuperCluster designs incorporate NVIDIA Quantum-2 InfiniBand or NVIDIA Spectrum-X Ethernet networking in a centralized rack, enabling a non-blocking, 256-GPU scalable unit in five racks or an extended 768-GPU scalable unit in nine racks.

All Supermicro NVIDIA HGX B200 systems are equipped with a 1:1 GPU-to-NIC ratio supporting NVIDIA BlueField-3 SuperNICs or NVIDIA ConnectX-7 NICs for scaling across a high-performance compute fabric, and natively support NVIDIA AI Enterprise software to accelerate time to production AI. As President and CEO Charles Liang stated, "In this transformative moment of AI, where scaling laws are pushing the limits of data center capabilities, our latest NVIDIA Blackwell-powered solutions, developed through close collaboration with NVIDIA, deliver outstanding computational power. Supermicro's NVIDIA Blackwell GPU offerings in plug-and-play scalable units with advanced liquid cooling and air cooling are empowering customers to deploy an infrastructure that supports increasingly complex AI workloads while maintaining exceptional efficiency." Supermicro provides support for systems included in the NVIDIA-Certified Systems program, which has certified Supermicro systems with NVIDIA H100 and H200 GPUs.