NVIDIA
NVIDIA L4 24GB
Single-slot low-profile 72 W accelerator with 24 GB GDDR6 for video pipelines, inference and virtual desktops.
| Availability | In stock |
|---|---|
| SKU | AXL-GPU-L4 |
| Minimum order | 1 unit |
| Condition | New |
| Pricing | By quotation |
Overview
The L4 is built for density rather than peak performance. A single-slot low-profile board drawing 72 W fits edge servers and standard 1U or 2U chassis without supplementary power cabling, so several cards can go into one node. Typical deployments are video encode and decode at scale, image and speech inference, recommendation serving and graphics-accelerated virtual desktops.
AXISLINK buys L4 through verified distribution and records the serial number of every board on the quotation. We supply the cards on their own, in tray quantities for integrators, or already installed and validated in a server platform sized for the card count you need. Packing, export documentation and freight are arranged from our Hong Kong office.
Specifications
| Architecture | NVIDIA Ada Lovelace |
|---|---|
| Memory | 24 GB GDDR6 |
| Memory bandwidth | 300 GB/s |
| Media engines | NVENC and NVDEC blocks with AV1 encode and decode |
| Form factor | PCIe 4.0 x16, single-slot low-profile, passively cooled |
|---|---|
| Display outputs | None |
| Virtualization | NVIDIA vGPU software supported |
| Board power | 72 W |
|---|---|
| Cooling | Passive, no supplementary power connector required |
Request a quote
Quotation for NVIDIA L4 24GB. We reply with a formal proforma invoice, typically within 24 hours.
More in GPUs & Accelerators
View category
NVIDIA
NVIDIA H200 141GB SXM
Hopper SXM module with 141 GB HBM3e for large-model training and memory-bound inference on HGX baseboards.
NVIDIA
NVIDIA L40S 48GB
Ada Lovelace data-center card with 48 GB GDDR6 ECC for inference, fine-tuning, rendering and virtual workstations.
NVIDIA
NVIDIA H100 80GB
Hopper-architecture data-center GPU for large-scale AI training and inference. PCIe and SXM variants.