GPUsNVIDIA · Ada, 24 GB, 72 W

L4Have or need NVIDIA L4 GPUs?

The NVIDIA L4 is a low-profile Ada Lovelace GPU with 24 GB of memory and a 72 W power envelope. It is built for efficient inference, video, and mainstream 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 L4, a low-profile single-slot card with a gold shroud, a bracket, and a PCIe connector
Image: NVIDIA
GPU memory
24GB
Memory bandwidth
300GB/s
Max TDP
72W
Architecture
Ada Lovelace

How the L4 compares

Per GPU, from the figures on each model's page.

GPU memory
Max power

What the L4 is

The L4 is NVIDIA's efficient inference and media card. It fits 1U and dense PCIe servers that cannot take a dual-slot 300–350 W board. People use it for video, inference, and general-purpose acceleration where power and space matter more than HBM capacity. It is not an SXM training GPU. If you have L40S or T4, use those pages.

Typical configs

L4 in a standard server

GPUs
1–8 × L4, 24 GB each
Interconnect
PCIe Gen4 x16, 64 GB/s
Media
2 NVENC, 4 NVDEC, 4 JPEG
CPU and memory
Dual socket, varies; often a 1U or 2U chassis
Power and rack
72 W per GPU, low-profile, single-slot. Eight cards are still a modest GPU load

What affects condition and lead time

  • Low-profile versus full-height brackets.
  • Hours, test records, and whether cards were used for 24/7 video encode.
  • How many cards per server, and whether the chassis ships with them.
  • Location. Every buyer and destination is screened against export-control and sanctions lists before the hardware moves.

Colo and power

  • 72 W per GPU. These are among the easiest data-center GPUs to land in an existing air-cooled row.
  • Confirm the server has the right slot length and low-profile clearance, not just the watts.
  • 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 L4?

A low-profile, single-slot Ada GPU with 24 GB of memory and a 72 W TDP, used for efficient inference, video, and mainstream servers.

How is the L4 different from the L40S and the T4?

L40S is a 48 GB, 350 W dual-slot card. T4 is the previous-generation 16 GB, 70 W low-profile card. L4 sits in the T4's power and size class, with 24 GB and Ada features.

Can I put L4s in a 1U server?

Yes, that is a common use. Confirm slot count, bracket type, and chassis airflow.

How much power does an L4 use?

72 W per card. An eight-card server is still a light GPU load compared with Hopper or Blackwell nodes.

How do I sell L4s 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 L4s 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 L4 GPUs?

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