GPUsNVIDIA · Turing, 16 GB, 70 W

T4Have or need NVIDIA T4 GPUs?

The NVIDIA T4 is a Turing inference GPU with 16 GB of GDDR6 memory and a 70 W power envelope, in a low-profile PCIe card. A lot of them are still coming out of production servers. Tell us what you have, or what you need. We verify the hardware, match both sides, and run the deal through payment and delivery. If you're selling, you pick which offer to take.

NVIDIA T4, a low-profile single-slot card with a silver shroud, a green stripe, and a bracket
Image: NVIDIA
GDDR6
16GB
Memory bandwidth
320GB/s
Max power
70W
Architecture
Turing

How the T4 compares

Per GPU, from the figures on each model's page. Where a model has more than one form, the form is named.

GPU memory
Max power

What the T4 is

The T4 (often still called Tesla T4) was NVIDIA's mainstream inference and virtualization card for a long stretch. It fits dense 1U servers. People still use it for inference, media, and VDI, or they sell it when they move a row to L4, L40S, or Hopper. If you have the newer L4, use that page.

Typical configs

T4 in a standard server

GPUs
1–8 × T4, 16 GB each (some chassis hold more in a dense 1U)
Interconnect
PCIe Gen3 x16
Power and rack
70 W per GPU, low-profile, passive. Needs chassis airflow, not a lot of PDU capacity

What affects condition and lead time

  • Hours and whether cards ran 24/7 inference.
  • Bracket type, and whether they are still in a server.
  • Driver and CUDA generation on the buyer's side — T4 is older than Ada and Hopper.
  • Location. Every buyer and destination is screened against export-control and sanctions lists before the hardware moves.

Colo and power

  • 70 W per card. These land in almost any air-cooled row that already hosted them.
  • The hardware does not include a colo site. If you need a place to land it, tell us the power, the cooling, and the site you have in mind.

The deal

How a deal works

How a used GPU deal works
  1. 01

    Verify

    Agreements first. We confirm the serials and condition, and qualify both sides.

  2. 02

    Match

    Vetted buyers make offers. The seller chooses which offer to take.

  3. 03

    Pay and deliver

    Every destination is screened. Payment happens with delivery.

Questions people ask

What is an NVIDIA T4?

A Turing data-center GPU with 16 GB of GDDR6 and a 70 W, low-profile PCIe design, widely used for inference and virtualization.

How is the T4 different from the L4?

L4 is Ada, with 24 GB and a 72 W low-profile board. T4 is the older 16 GB, 70 W Turing card. Same general size and power class, different generation.

Are T4s still used in production?

Yes. Many are still in service, and many are coming out as rows refresh. Tell us what you have or need and we will match it.

How much power does a T4 use?

70 W per card, passively cooled.

How do I sell T4s I have?

Tell us how many, the condition, and where they sit. We verify the hardware, match your Have with a Need, and handle the deal. You pick which offer to take.

How do I get T4s I need?

Tell us how many and where they will run. We match your Need with a Have and run the deal through payment and delivery. We don't publish prices.

How do you check buyers and export rules?

Every buyer and every seller passes a KYC identity and company check, and a Replenysh person reviews every buyer before they see a deal. Every buyer and destination is screened against export-control and sanctions lists before the hardware moves.

Have or need NVIDIA T4 GPUs?

Tell us what you have, or what you need. We verify, match, and handle the deal.