NVIDIA L40S

Cloud GPU Pricing — Live from 8+ Providers
Starting From
$0.600/hr
8 on-demand listings
VRAM
48 GB GDDR6
Architecture
Ada Lovelace
FP16 Performance
733 TFLOPS
Memory Bandwidth
864 GB/s
Listings
10 options
Provider Pricing
ProviderPrice / hrEst. MonthlyTypeVRAMAction
BestSalad Cloud $0.6000/hr $438/mo On-Demand 48 GB Rent Now →
Sesterce $0.7200/hr $525.6/mo On-Demand 48 GB Rent Now →
Hyperstack $0.8500/hr $620.5/mo On-Demand 48 GB Rent Now →
RunPod $0.8600/hr $627.8/mo On-Demand 48 GB Rent Now →
Verda $0.8900/hr $649.7/mo On-Demand 48 GB Rent Now →
Lambda Labs $0.9900/hr $722.7/mo On-Demand 48 GB Rent Now →
Vast.ai $1.0037/hr $732.7/mo On-Demand 45 GB Rent Now →
CoreWeave $1.2400/hr $905.2/mo On-Demand 48 GB Rent Now →
Vast.ai $0.4704/hr $343.39/mo Spot 45 GB Rent Now →
RunPod $0.7900/hr $576.7/mo Spot 48 GB Rent Now →
About the NVIDIA L40S

The NVIDIA L40S bridges consumer and data-center GPUs for inference workloads. With 48 GB GDDR6 and Ada Lovelace architecture, it excels at video generation, large diffusion models, and multi-modal inference at scale. More accessible than A100s while offering modern architecture efficiency gains.

GPU Specifications
VRAM48 GB GDDR6
ArchitectureAda Lovelace
FP16 Performance733 TFLOPS
Memory Bandwidth864 GB/s
TDP350W
InterconnectPCIe 4.0
Best For
Production InferenceVideo GenerationDiffusion ModelsMulti-modal Inference

Compare All NVIDIA L40S Options Side by Side

Filter by price type, sort by cost, and find the best deal for your workload.

Open Full Comparison →