L40S Cloud Rental Pricing
Ada Lovelace data-center GPU with 48GB, a cost-efficient choice for inference and image/video generation.
6 offers from 6 providers · updated October 4, 2026
Lowest on-demand
$0.47/hr
Median
$1.50/hr
Lowest spot
$0.78/hr
8x GPUs, 1 month
from $2,745
| Provider | Variant | VRAM | On-demand /hr | Spot /hr | Per month | Billing |
|---|---|---|---|---|---|---|
| Vast.ai | L40S | 48GB | $0.47 | — | $343 | per second |
| RunPod credits | L40S | 48GB | $1.09 | — | $796 | per ms |
| Crusoe | L40S | 48GB | $1.50 | — | $1,095 | per minute |
| Verda | L40S | 48GB | $1.56 | $0.78 | $1,139 | per hour |
| Vultr | L40S | 48GB | $1.67 | — | $1,219 | per hour |
| CoreWeave | L40S (min 8) | 48GB | $2.25 | $0.99 | $1,643 | per hour |
Per GPU-hour, on-demand unless noted. Monthly = 730 hours. Check each provider for region availability and multi-GPU node minimums.
L40S rental FAQ
How much does it cost to rent an L40S?
On-demand L40S rental ranges from $0.47 to $2.25 per GPU-hour across 6 providers (median $1.50), as of October 4, 2026. Spot/interruptible capacity starts at $0.78/hr.
What is the cheapest L40S cloud provider?
Vast.ai has the lowest on-demand L40S price we track at $0.47 per GPU-hour (L40S).
How much does an 8x L40S server cost per month?
At the cheapest on-demand rate ($0.47/GPU-hour), 8x L40S running 24/7 costs about $2,745 per month (730 hours). At the median rate it is about $8,760.