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
Silver 4U rackmount server chassis with front drive bays and rear redundant power supplies

ASUS

ASUS ESC8000A-E13-32W barebone server, dual EPYC 9005, 24 DDR5, 8 NVMe, 3200W PSU

SKU CH-9T3J-NJG5W MPN ESC8000A-E13-32W

4U barebone server housing dual AMD EPYC 9005 processors with 24 DDR5 ECC slots, eight hot-swap NVMe bays and 3+1 redundant 3200W Titanium PSU for eight dual-slot GPU accelerators

  • Supports eight dual-slot NVIDIA H200 or RTX GPUs for dense AI compute
  • Dual EPYC 9005 processors deliver up to 192 Zen 5c cores
  • 24 DDR5 ECC registered DIMMs across 12 channels reach 6400MHz
  • Eight hot-swappable NVMe bays provide high-speed storage flexibility
  • 3+1 redundant 3200W Titanium PSU ensures mission-critical power stability

Built for dense AI workloads

This 4U barebone server fits two AMD EPYC 9005 processors and up to eight dual-slot NVIDIA H200 or RTX accelerators. Twenty-four DDR5 ECC memory slots provide 12-channel bandwidth for large-model training and inference. A 3+1 redundant 3200W Titanium power supply keeps the chassis running under full GPU load.

Suits teams deploying large-scale training clusters

Organisations that need maximum GPU density in a single rack unit will match this platform. Buyers who require only one or two accelerators, or who prefer air-cooled single-socket designs, should look at smaller configurations.

The GPU capacity decides fit

Eight double-width cards at up to 600W each set the power and thermal envelope for the whole system. If the workload cannot use that many accelerators, the chassis and power supply become overspecified.

Highlights

  • Supports eight dual-slot NVIDIA H200 or RTX GPUs for dense AI compute
  • Dual EPYC 9005 processors deliver up to 192 Zen 5c cores
  • 24 DDR5 ECC registered DIMMs across 12 channels reach 6400MHz
  • Eight hot-swappable NVMe bays provide high-speed storage flexibility
  • 3+1 redundant 3200W Titanium PSU ensures mission-critical power stability

Specifications

No published specifications for this item yet.

Questions about this item

Should I pick the eight-GPU H200 setup or the RTX PRO 6000 and 4500 options for this 4U server?

The H200 and RTX PRO 6000 or 4500 Blackwell Server Edition GPUs all fit the eight dual-slot bays; choose H200 for maximum AI memory bandwidth or the RTX PRO models for mixed visualisation and compute workloads.

What does the 6400MHz DDR5 rating deliver once the 24 DIMM slots are populated?

Running all 24 DDR5 ECC registered modules at 6400MHz gives 12-channel bandwidth per CPU, keeping eight GPUs fed with sustained memory throughput for large-model training.

How does PCIe 5.0 limit the two expansion slots and the NVMe drives in practice?

The two PCIe 5.0 slots connect directly to the CPUs for low-latency NIC or DPU links, while the eight hot-swap bays support up to six 2.5" U.2 NVMe drives sharing the same Gen 5.0 lanes.

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.