We take cryptocurrency — so the coins you hold can buy real hardware.Pay in crypto, spend it on real hardware. Free insured shipping worldwide over $399. Plain, unbranded boxes, sent from Germany.Free insured shipping over $399. Your rate is locked for 30 minutes once checkout opens.Rate locked 30 minutes. Every item sealed, serial-checked and covered by a 24-month warranty.Sealed, 24-month warranty.

Saved items Sign in
5U rack server barebone with eight GPU modules

GIGABYTE

GIGABYTE G593-SD1-AAX3 barebone server 8 GPU 32 DIMM 900GB/s

SKU CH-17MR-48PT8 MPN G593-SD1-AAX3

A 5U barebone server supporting eight NVIDIA HGX H200 GPUs with 900GB/s interconnect bandwidth, dual Xeon processors and 32 DDR5 DIMM slots

  • NVIDIA HGX H200 8-GPU platform accelerates massive AI workloads
  • 900 GB/s GPU-to-GPU bandwidth via NVLink and NVSwitch cuts data bottlenecks
  • Dual Xeon Scalable processors deliver high compute density for HPC
  • 32 DDR5 RDIMM slots support large memory capacity per node
  • Dual 10 Gb/s LAN ports with Intel X710-AT2 enable fast cluster networking

HPC server with eight HGX H200 GPUs

The GIGABYTE G593-SD1-AAX3 is a 5U barebones platform built around the NVIDIA HGX H200 8-GPU baseboard. It targets organisations running large-scale AI training, high-performance computing and data-analytics workloads that demand massive parallel throughput.

900GB/s NVLink and NVSwitch interconnect

The system delivers 900GB/s GPU-to-GPU bandwidth through NVIDIA NVLink and NVSwitch, removing the PCIe bottleneck that limits multi-GPU scaling in standard servers. The jump is worth it when model size or dataset throughput exceeds what a single GPU or slower interconnect can sustain.

Dual Xeon CPUs and 32 DDR5 RDIMM slots

Two 5th or 4th Gen Intel Xeon Scalable processors, including Xeon CPU Max Series, drive 8-channel DDR5 memory across 32 DIMM slots. Dual ROM architecture adds firmware resilience, while two 10Gb/s LAN ports via Intel X710-AT2 handle management and data traffic. Compatibility with NVIDIA BlueField-3 DPUs and SuperNICs offloads networking and storage tasks from the GPUs.

Highlights

  • NVIDIA HGX H200 8-GPU platform accelerates massive AI workloads
  • 900 GB/s GPU-to-GPU bandwidth via NVLink and NVSwitch cuts data bottlenecks
  • Dual Xeon Scalable processors deliver high compute density for HPC
  • 32 DDR5 RDIMM slots support large memory capacity per node
  • Dual 10 Gb/s LAN ports with Intel X710-AT2 enable fast cluster networking

Specifications

No published specifications for this item yet.

Questions about this item

What rack height does the 5U chassis need?

The server occupies a standard 5U rack space; verify your cabinet has at least five contiguous rack units and sufficient depth for the 8-GPU HGX H200 board.

How are the dual Intel Xeon processors and 32 DDR5 RDIMMs installed, and what step is often missed?

Install both CPUs first, then populate all 32 DDR5 RDIMM slots evenly across the eight memory channels; the step most often missed is seating every DIMM in the correct channel order for balanced bandwidth.

Which field-replaceable parts will need attention after extended GPU compute runs?

The hot-swap power supplies and system fans are the primary field-replaceable units; monitor their health LEDs and replace any module that reports a fault to maintain cooling for the eight HGX H200 GPUs.

New to this? Buying server hardware with crypto — what to check, what it costs to pay, and what is never asked for.

Also in this aisle

People compared these

Same shelf, same checkout — eight coins and a 30-minute rate lock.