Brand: Gigabyte | Category: GPUs
SKU: GV-NL40S48-00 | Part #: GV-NL40S48-00 | MPN: GV-NL40S48-00
Contact for Pricing — Request a Quote
Gigabyte NVIDIA L40S 48GB PCIe GPU Accelerator Card is a Gigabyte-engineered enterprise GPU designed for AI inference, professional visualization, and virtual desktop infrastructure (VDI) deployments at scale. Built on NVIDIA's Ada architecture, this dual-slot accelerator delivers exceptional performance-per-watt for mission-critical workloads.
Key Features:
Perfect for:
| Brand | Gigabyte |
| Category | GPUs |
| SKU | GV-NL40S48-00 |
| Part Number | GV-NL40S48-00 |
| Condition | New |
| Form Factor | PCIe GPU Accelerator Card |
| GPU Memory | 48 GB GDDR6 |
| GPU Architecture | NVIDIA Ada Lovelace |
| CUDA Cores | 14,080 |
| Tensor Performance | 1.457 PetaFLOPS (FP8) |
| PCIe Interface | PCIe 5.0 x18 |
| Video Output | Not applicable (compute-only) |
| Power Consumption | 350W TDP |
| Power Connectors | 2x 8-pin PCIe power |
| Length | 267.4 mm |
| Height | 111.15 mm |
| NVIDIA Compute Capability | 8.9 |
The Gigabyte NVIDIA L40S 48GB PCIe GPU Accelerator Card accelerates AI/ML training and inference, scientific HPC and virtualised GPU workloads in data-centre servers. Typical deployments include model training clusters, inference serving, computer-vision pipelines and rendering farms.
Key specifications for the Gigabyte NVIDIA L40S 48GB PCIe GPU Accelerator Card: new condition; form factor PCIe GPU Accelerator Card; gpu memory 48 GB GDDR6; gpu architecture NVIDIA Ada Lovelace; cuda cores 14,080; tensor performance 1.457 PetaFLOPS (FP8); pcie interface PCIe 5.0 x18. Manufacturer part number GV-NL40S48-00. For the full datasheet with electrical, environmental, and compliance details, contact our pre-sales engineering team.
The Gigabyte NVIDIA L40S 48GB PCIe GPU Accelerator Card needs a GPU-qualified server with adequate power and cooling for its TDP, plus NVIDIA CUDA driver support in your hypervisor or bare-metal OS. Our team confirms chassis qualification, slot or baseboard type and PSU headroom before quoting.