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
Full-height full-length dual-slot passive AI accelerator card with 94 GB HBM3 memory

Supermicro

Supermicro GPU-NVH100NVL H100 NVL 94 GB HBM3 ai accelerator card

SKU CH-KXES-CRFAQ MPN GPU-NVH100NVL

Delivers 94 GB HBM3 memory and Tensor acceleration for AI workloads via PCIe 5.0 x16 in a dual-slot fanless design

  • InterfacePCI Express 5.0 x16
  • Chipset ManufacturerNVIDIA
  • GPUH100 Tensor Core GPU
  • Core ClockBase: 1080 MHz Boost: 1785 MHz
  • Memory Size94GB
  • Memory Clock2619 MHz
  • 94 GB HBM3 memory handles massive AI models
  • PCI Express 5.0 x16 interface delivers high bandwidth
  • H100 Tensor Core GPU accelerates training and inference
  • Fanless design enables dense server deployments
  • 400 W default power mode balances performance and efficiency

H100 NVL AI accelerator with 94 GB HBM3

The Supermicro GPU-NVH100NVL is an NVIDIA H100 NVL Tensor Core GPU card built for data-centre AI training and inference workloads. It connects through a PCI Express 5.0 x16 interface and provides 94 GB of HBM3 memory for large-model processing. This full-height, full-length dual-slot card targets server integrators deploying high-density compute nodes.

Passive cooling with configurable power modes

The card uses a fanless heatsink and relies on chassis airflow within an ambient range of 0°C to 50°C for sustained operation. Power delivery uses a 16-pin auxiliary connector that can be strapped for different limits: a 450 W or 600 W mode allows a 400 W maximum with a 310 W compliance limit, while a 300 W mode caps draw at 310 W. The GPU core clocks run at a 1080 MHz base and 1785 MHz boost under these thermal constraints.

Supported operating systems and power cabling

Driver support covers Windows 10, Windows 11, Windows Server 2019 and Windows Server 2022 through WHQL certification. The only bundled hardware is the 16-pin power cable, which must be strapped to match the chosen power envelope. No additional software utilities, mounting brackets or active cooling accessories are included with the card.

Highlights

  • 94 GB HBM3 memory handles massive AI models
  • PCI Express 5.0 x16 interface delivers high bandwidth
  • H100 Tensor Core GPU accelerates training and inference
  • Fanless design enables dense server deployments
  • 400 W default power mode balances performance and efficiency

Specifications

BrandNVIDIA
ModelH100 NVL
Part Number900-21010-0020-000
InterfacePCI Express 5.0 x16
Chipset ManufacturerNVIDIA
GPUH100 Tensor Core GPU
Core ClockBase: 1080 MHz
Boost: 1785 MHz
Memory Size94GB
Memory Clock2619 MHz
Memory TypeHBM3
CoolerFanless
AI FeaturesAI Accelerated
Operating Systems SupportedWindows Hardware Quality Lab (WHQL):
- Windows 10, Windows 11
- Windows Server 2019, Windows Server 2022
Auxiliary Power Connector16-pin auxiliary power connector
PowerPCIe 16-pin cable strapped for 450 W or 600 W power mode:
> 400 W maximum (default)
> 310 W power compliance limit1
> 200 W minimum
PCIe 16-pin cable strapped for 300 W power mode:
> 310 W maximum (default)
> 310 W power compliance limit
> 200 W minimum
Form FactorFull-height, full-length (FHFL) 10.5"
Slot WidthDual-slot
EnvironmentAmbient operating temperature: 0°C to 50°C
Ambient operating temperature (short term): -5°C to 55°C
Storage temperature: -40°C to 75°C
Operating humidity (short term): 5% to 93% relative humidity
Operating humidity: 5% to 85% relative humidity
Storage humidity: 5% to 95% relative humidity
Shipping weight1.54 kg
Package size345 × 260 × 83 mm

Questions about this item

How do I install the Supermicro H100 NVL 94 GB HBM3 card in a PCIe 5.0 x16 slot and connect the 16-pin auxiliary power cable?

Insert the full-height, full-length dual-slot card into a PCIe 5.0 x16 slot, secure the bracket, then attach the 16-pin auxiliary power connector; the cable must be strapped for the desired power mode (450 W/600 W or 300 W) before powering on.

What maintenance will the fanless 94 GB HBM3 accelerator need during its service life?

The passive cooler has no moving parts; schedule periodic dust removal from the heatsink fins and verify that chassis airflow keeps the ambient temperature within the 0°C to 50°C operating range.

Which thermal or power limit is reached first when the H100 Tensor Core GPU runs sustained workloads at the default 400 W setting?

The 310 W power compliance limit is enforced before the 400 W maximum, so sustained loads will throttle at the compliance ceiling rather than the higher default ceiling.

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.