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 PCIe card, silver metal bracket, black PCB, 16 GB GDDR6

NVIDIA

NVIDIA 900-2G183-0000-001 Tesla T4 16 GB GDDR6 ai accelerator card

SKU CH-M8EA-75WRJ MPN 900-2G183-0000-001

16 GB GDDR6 AI accelerator with multi-precision Tensor cores in a 70-watt PCIe form factor

  • 16 GB GDDR6 memory powers large-model inference
  • 70 watt design fits dense scale-out servers
  • Turing Tensor Cores accelerate mixed-precision workloads
  • Full-height PCIe form factor slots into standard racks
  • New condition ensures full factory specification

AI accelerator for scale-out servers

The NVIDIA Tesla T4 is a 16 GB GDDR6 AI accelerator card built for scale-out computing environments. It targets data-centre workloads such as deep learning training, inference, machine learning, data analytics and graphics. The full-height PCIe form factor fits standard server slots. Buyers who need more than 16 GB of memory or multiple cards per node should look elsewhere.

70-watt thermal envelope

The card operates within a 70-watt power envelope, enabling energy-efficient deployment in dense racks. Its small PCIe form factor is optimised for scale-out servers where thermal headroom is limited. No external power connectors are required. Installations that cannot accommodate a full-height card or that demand lower power draw will need a different solution.

NGC software stack integration

Accelerated containerised software stacks from NGC are combined with the hardware to deliver performance at scale. The bundled software simplifies deployment of diverse cloud workloads. Users who prefer to manage their own software stack or who rely on unsupported frameworks will not benefit from this integration.

Highlights

  • 16 GB GDDR6 memory powers large-model inference
  • 70 watt design fits dense scale-out servers
  • Turing Tensor Cores accelerate mixed-precision workloads
  • Full-height PCIe form factor slots into standard racks
  • New condition ensures full factory specification

Specifications

No published specifications for this item yet.

Questions about this item

How is the Tesla T4 16 GB GDDR6 card installed in a server?

Insert the full-height PCIe card into a compatible x16 slot, secure the bracket screw, and ensure the chassis airflow meets the 70-watt thermal design; the step often missed is verifying that the server BIOS allocates enough PCIe resources for the 16 GB frame buffer.

What maintenance or replacement will the Tesla T4 16 GB GDDR6 card need later?

Plan to clean the passive heatsink fins annually and replace the card if the 16 GB GDDR6 memory or Tensor cores show persistent ECC errors; no fan or liquid-loop service is required because the 70-watt design uses passive cooling only.

What limit appears first on the Tesla T4 16 GB GDDR6 card under sustained AI workloads?

The 70-watt power envelope is the first ceiling; once the Tensor cores drive the GPU to that sustained draw, the card throttles clocks to stay within the thermal budget before the 16 GB GDDR6 capacity becomes a bottleneck.

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.