Ada Lovelace In stock

NVIDIA L40S

Ada-generation inference and rendering with hardware video encode.

Best suited to

Inference serving
Video transcoding
3D rendering and Omniverse

Available regions

Region	Location	Typical latency
us-east-1	US East (Ashburn, VA)	12 ms
us-west-2	US West (Hillsboro, OR)	38 ms
eu-west-1	EU West (Dublin, IE)	84 ms

Hourly rate

$0.99/hr

Billed per second. $23.76 for a full day, $723 if left running a month.

Deploy this GPU All pricing

Full specification

ArchitectureAda Lovelace
VRAM48 GB
Memory typeGDDR6 ECC
Memory bandwidth864 GB/s
FP16 dense733 TFLOPS
FP8 dense1,466 TFLOPS
CUDA cores18,176
Tensor cores568 (4th gen)
Board power350W
InterconnectNot supported
Host vCPU16
Host RAM128 GB
NVMe scratch800 GB
Network200 Gbps Ethernet
MIG partitioningNo

Other Ada Lovelace cards

RTX 4090

24 GB GDDR6X