The market is changing daily. Stay on top of changes with our June Market Update »
Servers
Back
New Dell PowerEdge Servers
Back
Dell PowerEdge 17th Generation
Dell PowerEdge 16th Generation
Refurbished Dell PowerEdge Servers
Back
Dell PowerEdge 15th Generation
Dell PowerEdge 14th Generation
New HPE ProLiant Servers
Back
HPE ProLiant Gen12
HPE ProLiant Gen11
HPE ProLiant Gen10
Refurbished HPE ProLiant Servers
Back
HPE ProLiant Gen10 Plus
HPE ProLiant Gen10
AI/ML Servers
Back
Supermicro GPU Servers
HPE GPU Servers
Dell EMC GPU servers
New Supermicro Servers
Back
SuperMicro X14 Servers
SuperMicro GPU Servers
SuperMicro AMD Servers
Towers & Workstations
Back
Dell 16th Gen Tower Servers
Dell 15th Gen Tower Servers
Dell 14th Gen Tower Servers
Dell Workstations
HPE Towers
ServerMonkey Featured Deals
Back
Refurbished Dell PowerEdge R760 12x3.55''
Refurbished Dell PowerEdge R640 8x2.5''
Refurbished Dell PowerEdge R650 10x2.5''
Refurbished HPE ProLiant DL360 Gen10 8x2.5''
Storage
Back
Dell EMC Storage
Back
Dell PowerVault
Dell PowerStore
Dell PowerMax
Dell PowerScale
HPE Storage
Back
HPE MSA Shared Storage
HPE Alletra
Synology Storage
Back
FS Series
RS Series
SA Series
UC Series
QNAP Storage
Back
Enterprise
Small/Medium Business (SMB)
Home
Expansion Units
Networking
Back
Dell Networking
Back
Refurbished Dell Networking
New Dell Networking
HPE Networking
Back
FlexNetwork Series Switches
FlexFabric Series Switches
Cisco Networking
Back
IP Phones
Routers
Switches
Wireless
Aruba Networking
Back
Access Switches
Campus Core And Data Center Switches
Enterprise Wireless LAN
Juniper Networking
Back
Firewall
Wireless Access Points and Edge
Switches
Routers
Meraki Networking
Back
Meraki Switches
Access Points
Arista Networking
Back
7800 Series
7500 Series
7300 Series
7280 Series
7160 Series
7150 Series
7130 Series
7060X Series
7050X Series
7020R Series
7010 Series
750 Series
720X Series
Fortinet Networking
Back
Firewalls
SD-WAN
Secure Ethernet Switches
Secure Wireless LAN
Ubiquiti Networking
Back
Switches
WiFi
Sophos Networking
Back
Firewalls
SD-WAN
WiFi
SonicWall Networking
Back
Firewall
Wireless Access Points
Switches
Extreme Networking
Back
Switches
Wireless Access Points
Routing
Components
Back
Parts & Upgrades
Back
Processors
Memory
Drives
Network Cards & Adapters
RAID & Storage Adapters
Other Parts & Upgrades
Optics
Back
Dell Optics
HPE Optics
Accelerators
Back
Nvidia
AMD
Intel
Maintenance
ITAD
Back
IT Asset Disposition
Buy Back
Sustainability
Contact Us
Introduction

The Most Powerful Universal GPU

The NVIDIA L40S delivers breakthrough multi-workload performance for the data center. Combining powerful AI compute with best-in-class graphics and media acceleration, the L40S is built to power generative AI, LLM inference and training, 3D rendering, real-time ray tracing, and virtual desktop infrastructure. Based on the Ada Lovelace architecture with fourth-generation Tensor Cores, third-generation RT Cores, and 48GB of GDDR6 memory, the L40S is the ideal universal accelerator for enterprises running diverse workloads on a single GPU platform.

Powered by Ada Lovelace Architecture

The Ada Lovelace architecture delivers fourth-generation Tensor Cores with FP8 support, third-generation RT Cores for real-time ray tracing, a Transformer Engine for accelerated LLM performance, and DLSS 3 for AI-enhanced frame generation. All in a 350W, dual-slot, air-cooled PCIe card.

Learn About Ada Lovelace
Highlights

Universal Performance

1,466TFLOPS FP8 Tensor
AI Compute Performance
212TFLOPS RT Core
Real-Time Ray Tracing
91.6TFLOPS FP32
Single-Precision Compute
Workloads

Multi-Workload Acceleration

Generative AI

Generative AI

With next-generation AI, graphics, and media acceleration capabilities, the L40S delivers up to 5X higher inference performance than the previous-generation NVIDIA A40. With 48GB of memory capacity and FP8 Tensor Core support, the L40S is the ideal platform for accelerating multimodal generative AI workloads, from image generation to content creation.

Learn More About Generative AI
LLM Training & Inference

LLM Training & Inference

Fourth-generation Tensor Cores with FP8 support deliver exceptional AI computing performance for training and inference on state-of-the-art large language models. The Transformer Engine intelligently manages precision across transformer layers, automatically recasting between FP8 and FP16 to maximize speed without sacrificing accuracy.

Explore NVIDIA's Inference Platform
Rendering & 3D Graphics

Rendering & 3D Graphics

Third-generation RT Cores deliver up to 2X the real-time ray-tracing performance of the previous generation, powering the creation of stunning visual content and high-fidelity creative workflows. DLSS 3 frame generation leverages deep learning and the Optical Flow Accelerator to boost rendering performance and deliver higher frame rates with lower latency.

Learn More About NVIDIA RTX
VDI & Enterprise Data Center

VDI & Enterprise Data Center

The L40S is optimized for 24/7 enterprise data center operations. At 350W with passive cooling and PCIe Gen4 interface, it fits into standard rack infrastructure without liquid cooling. NEBS Level 3 ready, secure boot with root of trust, and vGPU software support make it enterprise-ready for virtual desktop infrastructure and multi-tenant environments.

Learn More About MIG
Server Compatibility

Compatible Dell PowerEdge Servers

Compatible Dell PowerEdge Servers

Dell PowerEdge R760xa

Dell PowerEdge R760xa

16th Gen
  • Versatile 2U GPU server
  • Supports L40S for AI and graphics
View Server
Dell PowerEdge R7625

Dell PowerEdge R7625

16th Gen
  • AMD EPYC powered 2U rack server
  • Multi-GPU capable for L40S
View Server
Dell PowerEdge R7725

Dell PowerEdge R7725

17th Gen
  • Latest-gen AMD EPYC server
  • L40S compatible for AI and VDI
View Server
Dell PowerEdge XE7745

Dell PowerEdge XE7745

17th Gen
  • High-density GPU platform
  • Optimized for multi-GPU L40S
View Server
Dell PowerEdge R760

Dell PowerEdge R760

16th Gen
  • Mainstream 2U enterprise server
  • L40S for mixed AI/graphics loads
View Server
Dell PowerEdge R570

Dell PowerEdge R570

17th Gen
  • Compact 17th Gen rack server
  • L40S for edge and enterprise AI
View Server
View Full GPU Compatibility Matrix
Specifications

NVIDIA L40S GPU

Specification L40S
GPU Architecture NVIDIA Ada Lovelace
GPU Memory 48GB GDDR6 with ECC
Memory Bandwidth 864 GB/s
Interconnect PCIe Gen4 x16 (64 GB/s bidirectional)
CUDA Cores 18,176
RT Cores (3rd Gen) 142
Tensor Cores (4th Gen) 568
RT Core Performance 212 TFLOPS
FP32 91.6 TFLOPS
TF32 Tensor Core 183 / 366 TFLOPS*
BFLOAT16 Tensor Core 362 / 733 TFLOPS*
FP16 Tensor Core 362 / 733 TFLOPS*
FP8 Tensor Core 733 / 1,466 TFLOPS*
INT8 Tensor 733 / 1,466 TOPS*
Form Factor 4.4" H × 10.5" L, dual slot
Display Ports 4× DisplayPort 1.4a
Max Power 350W
Power Connector 16-pin
Thermal Passive
vGPU Support Yes
NVENC / NVDEC 3× / 3× (includes AV1 encode/decode)
Secure Boot Yes (Root of Trust)
NEBS Ready Level 3
MIG Support No
NVLink Support No
* Dense / Sparse. Sparse performance uses structural sparsity.
View Datasheet View Product Brief
 

Interested in This GPU?

Get pricing, availability, and compatibility details from the ServerMonkey team.

Request a Quote

What are you looking for?

Need help?