NVIDIA DGX B200 GPU Server

Brand: NVIDIA | Category: Servers

SKU: NV-DGX-B200 | Part #: DGX-B200 | MPN: DGX-B200

Contact for Pricing — Request a Quote

The NVIDIA DGX-B200 is a 2U rack-mount GPU-accelerated server designed for next-generation artificial intelligence, high-performance computing, and large language model workloads. Built on the EIA-310 19-inch rack standard, this system houses multiple NVIDIA B200 Tensor GPUs, each equipped with 192 GB of memory, delivering exceptional compute density for enterprise data centers. The DGX-B200 accommodates either Intel Xeon or AMD EPYC host processors, allowing organizations to standardize on their preferred CPU architecture while maintaining access to NVIDIA's latest GPU acceleration technology.

AI infrastructure teams, data center architects, and machine learning operations engineers select the DGX-B200 to consolidate training and inference pipelines without sacrificing performance or reliability. The system's hot-swap NVMe, SAS, and SATA drive configurations adapt to varied storage tiers, while redundant hot-swap power supplies and dual cooling modules—liquid and air-cooled—ensure continuous uptime in demanding production environments. Management integration via IPMI 2.0, vendor-specific iDRAC/iLO/XCC interfaces, and the NVIDIA DGX Management Console simplifies fleet operations at scale. Configurable interconnect options including NVLink, PCIe 5.0, and 400GbE InfiniBand enable seamless multi-GPU scaling and high-speed cluster communication.