(858) 613-1818Quote list 0
Available on request

8 GPU HGX H200 AI server

Eight NVIDIA H200 SXM GPUs with 141 GB HBM3e each, NVLink and NVSwitch, 800 or 400 Gb/s fabric ports, in an 8U air cooled chassis, on request.

GPU: NVIDIA HGX H200

Illustration of an 8U GPU server
GPUs
8 x NVIDIA H200 SXM, 141 GB HBM3e each, NVLink and NVSwitch
CPU
2 x AMD EPYC 9005 or 2 x Intel Xeon 6
Memory
Up to 3 TB DDR5 RDIMM
Storage
2 x M.2 boot, up to 8 x U.2 NVMe
Networking
8 x 800 Gb/s ConnectX-8 or 8 x 400 Gb/s ConnectX-7 (east west), 2 x 400 Gb/s frontend
Power
Redundant PSUs, about 10 kW per node planning figure

Overview

The HGX H200 node is the Hopper generation eight GPU server, still widely deployed for training and inference and attractive for teams standardized on Hopper. Each H200 carries 141 GB of HBM3e at 4.8 TB/s and the eight GPUs are linked by fourth generation NVLink and NVSwitch at 900 GB/s per GPU, giving the node 1.1 TB of GPU memory.

We list it as available on request because supply now follows the Blackwell ramp and configurations vary by batch. It suits expansion of an existing H100 or H200 cluster, workloads validated on Hopper that will not move soon, and budgets that favor a proven platform over the newest one. The lower planning power of about 10 kW per node also helps in halls with limited rack density.

Host options are the same as the Blackwell classes: dual AMD EPYC 9005 or Intel Xeon 6, DDR5 to 3 TB, mirrored boot and up to eight U.2 drives. Fabric is 400 Gb/s ConnectX-7 for NDR InfiniBand and 400 GbE sites, or 800 Gb/s ConnectX-8 where the node will join a Quantum-X800 or Spectrum-X 800G fabric.

Use cases

  • Expanding Hopper clusters
  • Inference on validated Hopper stacks
  • Fine tuning and RAG pipelines
  • Research clusters

Specifications

Compute
GPUs8 x NVIDIA H200 SXM on the HGX H200 baseboard
GPU memory141 GB HBM3e per GPU, 1.1 TB per node, 4.8 TB/s per GPU
GPU interconnectFourth generation NVLink, 900 GB/s per GPU, NVSwitch
GPU powerUp to 700 W per GPU
CPUs2 x AMD EPYC 9005 series or 2 x Intel Xeon 6
System memory24 or 32 x DDR5 RDIMM, up to 3 TB
Storage
Boot2 x M.2 NVMe, mirrored
DataUp to 8 x U.2 NVMe, 7.68 TB to 30.72 TB each
RAIDSoftware RAID or optional hardware RAID for boot volume
Networking
GPU fabric8 x 800 Gb/s ConnectX-8 or 8 x 400 Gb/s ConnectX-7, InfiniBand or RoCE
Frontend2 x 400 Gb/s, BlueField-3 or ConnectX-7
ManagementDedicated 1 GbE BMC port, TPM 2.0
Physical
Form factor8U rackmount, air cooled
Power suppliesRedundant 3,000 W class PSUs, 200 to 240 V input
Planning powerAbout 10 kW per node

Figures describe the system class. Exact configuration, weights and power draw are confirmed on the quotation.

What's included

  • Rails, power cords and cable management arm
  • Burn in with multi hour GPU stress and NVLink validation, acceptance report
  • NVIDIA end user registration handled with the order
  • Operating system imaging, driver and fabric manager baseline on request
  • 3 year warranty, options to 5 years with advance replacement

Compliance

Compare it with the next class up or down

The nearest classes we quote are the 8 GPU HGX B200 AI server and the 8 GPU H200 NVL PCIe AI server. If neither is quite right, tell us the workload and we specify the system between them.

Talk to an engineerRequest a quotation