NVIDIA · Ada Lovelace · L-series
48 GB of GDDR6 at 0.9 TB/s, built for inference, fine-tuning, rendering.
From $1.19/hracross 7 providers
Memory
48 GB
GDDR6
Bandwidth
0.86 TB/s
FP16 tensor
362 TFLOPS
FP8 tensor
733 TFLOPS
Interconnect
PCIe
TDP
350 W
CUDA cores
18,176
Tensor cores
568
| Provider | Type | Spot | Details |
|---|---|---|---|
 RunPod | neocloud | Yes | View |
 CoreWeave | neocloud | — | View |
 Nebius | neocloud | — | View |
 Crusoe | neocloud | — | View |
 Amazon AWS | hyperscaler | Yes | View |
 Google Cloud | hyperscaler | Yes | View |
 Oracle Cloud | hyperscaler | Yes | View |
