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
7U rackmount server chassis with eight GPU modules installed

ASUS

ASUS HGX 8x H200 GPU Server barebone server

SKU CH-SV56-ENAZT MPN HGX 8x H200 GPU Server

Eight-GPU barebones server with dual Xeon processors and 900GB/s interconnect for large-scale AI and HPC workloads

  • Dual 5th Gen Intel Xeon processors support 350W TDP for heavy workloads
  • NVLink interconnect delivers 900GB/s bandwidth for efficient multi-GPU scaling
  • One-GPU-to-one-NIC topology supports up to 12 NICs for maximum throughput
  • Modular design reduces cable usage to shorten assembly and improve thermal flow
  • 4+2 80 PLUS Titanium power supplies provide high power efficiency

Eight-GPU AI server barebones

This ASUS HGX system is a 7U barebones server built for large-scale AI and high-performance computing. It integrates eight NVIDIA H200 GPUs with dual 5th Gen Intel Xeon Scalable processors. The platform targets organisations training massive generative models or running dense HPC simulations. It is not suited for buyers who need a compact tower or a general-purpose file server.

Direct-to-chip cooling with Titanium power

The thermal design uses dedicated CPU and GPU airflow tunnels plus support for direct-to-chip liquid cooling to handle extreme heat loads. Four-plus-two 80 PLUS Titanium power supplies deliver high-efficiency conversion for the dual 350W TDP processors and eight accelerators. A modular layout reduces internal cabling to improve airflow paths. Buyers without facility liquid-cooling infrastructure cannot use the D2C option.

One-GPU-to-one-NIC topology for scaling

A dedicated one-GPU-to-one-NIC architecture supports up to eight network interfaces for maximum throughput during compute-intensive workloads. GPU-to-GPU interconnect via NVLink provides 900GB/s bandwidth for efficient multi-GPU scaling. The chassis offers up to 12 PCIe slots for additional expansion cards. Organisations needing only single-GPU inference will find the networking topology excessive.

Highlights

  • Dual 5th Gen Intel Xeon processors support 350W TDP for heavy workloads
  • NVLink interconnect delivers 900GB/s bandwidth for efficient multi-GPU scaling
  • One-GPU-to-one-NIC topology supports up to 12 NICs for maximum throughput
  • Modular design reduces cable usage to shorten assembly and improve thermal flow
  • 4+2 80 PLUS Titanium power supplies provide high power efficiency

Specifications

No published specifications for this item yet.

Questions about this item

How do I install the direct-to-chip cooling loop and what step is most often missed?

Route the cold-plate manifolds through the dedicated CPU and GPU airflow tunnels, then pressure-test every quick-connect fitting before powering on; the step most often missed is torquing each fitting to spec after the first thermal cycle.

Which components will need scheduled replacement during the server's service life?

The 4+2 80 PLUS Titanium power supplies are the primary field-replaceable units; plan on swapping any supply that exceeds its rated MTBF or shows efficiency drift.

What thermal limit appears first when all eight GPUs run sustained workloads?

The 350 W TDP per processor sets the ceiling; once combined CPU and GPU heat load saturates the direct-to-chip loop capacity, temperatures climb and throttling begins.

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.