Skip to content

NVIDIA L40S

An Ada-generation datacenter card for inference, rendering, and mixed graphics workloads.

The lowest price you can rent right now is $1.09/GPU/hr, at RunPod in OC-AU-1.

Cheapest L40S in stock right now

As of Nov 16, 22:30 UTC

$1.09/GPU/hr

Same as the 7-day median

  • Low $1.09
  • Median $1.09
  • High $1.65
  • In stock 83% of hours
Line chart with 168 data points. Use the left and right arrow keys to move between data points, Page Up and Page Down to jump by a tenth of the data, Home and End to jump to the first and last, Enter or Space to activate the current point, and Escape to dismiss the readout.

$1.09 at RunPod was in stock 61% of the last 7 days

Rent it at RunPod L40S in OC-AU-1
  • Ada Lovelace
  • 48 GB GDDR6
  • 4 providers

The lowest L40S price is one you can get: $1.09/GPU/hr.

RunPod's $1.09/GPU/hr instance is listed as in stock, but the provider doesn't say how often.

How we pick these prices · What availability means

Provider by provider

Each provider's lowest on-demand L40S price, and how often that instance was in stock over the last 30 days.

ProviderLowest priceInstanceRegionIn stock, last 30 daysLowest in stock at last scan
RunPod$1.09/GPU/hrL40SEU-NL-1Only reports in stock$1.09/GPU/hr
DigitalOcean$1.57/GPU/hrL40S GPU Droplet - 1Xtor11%Sold out
Scaleway$1.62/GPU/hrL40S-1-48Gfr-par-269%$1.62/GPU/hr
AWS$1.86/GPU/hrg6e.xlargeus-east-1Not publishedNot published
Oracle Cloud$3.50/GPU/hrBM.GPU.L40S.4globalNot publishedNot published

AWS and Oracle Cloud list it from $1.86/GPU/hr, 1.7× the lowest price here. That's the published on-demand rate, not a scan error: reserved capacity costs less, and spot starts at $0.56/GPU/hr at AWS. Neither reports stock.

How often each provider had one in stock

Line chart with 719 data points. Use the left and right arrow keys to move between data points, Page Up and Page Down to jump by a tenth of the data, Home and End to jump to the first and last, Enter or Space to activate the current point, and Escape to dismiss the readout.

What the NVIDIA L40S is

An Ada Lovelace datacenter card with 48 GB, built from the same chip as the RTX 6000 Ada. It runs FP8, has no NVLink, and is sold as inference and graphics capacity rather than for large training jobs.

Architecture
Ada Lovelace
Memory
48 GB GDDR6
Memory bandwidth
0.864 TB/s
BF16 compute
362 TFLOPS, dense
Lowest precision
FP8

Other GPUs to consider

Head to head:L40S vs A100L40S vs RTX 6000 Ada

104 NVIDIA L40S instances

Availability
L40SRunPod · OC-AU-1
$1.09/GPU/hr
—
available
L40S-1-48GScaleway · fr-par-2
$1.62/GPU/hr
—
available
g6e.xlargeAWS · us-east-1
$1.86/GPU/hr
—
not reported
g6e.xlargeAWS · us-east-2
$1.86/GPU/hr
—
not reported
g6e.xlargeAWS · us-west-2
$1.86/GPU/hr
—
not reported
g6e.xlargeAWS · eu-north-1
$1.97/GPU/hr
—
not reported
g6e.xlargeAWS · ap-south-1
$2.24/GPU/hr
—
not reported
g6e.2xlargeAWS · us-east-1
$2.24/GPU/hr
—
not reported
g6e.2xlargeAWS · us-east-2
$2.24/GPU/hr
—
not reported
g6e.2xlargeAWS · us-west-2
$2.24/GPU/hr
—
not reported
g6e.xlargeAWS · ap-northeast-2
$2.29/GPU/hr
—
not reported
g6e.xlargeAWS · eu-central-1
$2.33/GPU/hr
—
not reported
g6e.2xlargeAWS · eu-north-1
$2.38/GPU/hr
—
not reported
g6e.12xlargeAWS · us-east-1
$2.62/GPU/hr
$10.49/hr total
not reported
g6e.12xlargeAWS · us-east-2
$2.62/GPU/hr
$10.49/hr total
not reported
g6e.12xlargeAWS · us-west-2
$2.62/GPU/hr
$10.49/hr total
not reported
g6e.2xlargeAWS · ap-south-1
$2.69/GPU/hr
—
not reported
g6e.xlargeAWS · ap-northeast-1
$2.70/GPU/hr
—
not reported
g6e.xlargeAWS · ap-northeast-3
$2.70/GPU/hr
—
not reported
g6e.2xlargeAWS · ap-northeast-2
$2.76/GPU/hr
—
not reported
g6e.12xlargeAWS · eu-north-1
$2.78/GPU/hr
$11.13/hr total
not reported
g6e.2xlargeAWS · eu-central-1
$2.80/GPU/hr
—
not reported
g6e.4xlargeAWS · us-east-1
$3.00/GPU/hr
—
not reported
g6e.4xlargeAWS · us-east-2
$3.00/GPU/hr
—
not reported
g6e.4xlargeAWS · us-west-2
$3.00/GPU/hr
—
not reported
g6e.12xlargeAWS · ap-south-1
$3.15/GPU/hr
$12.60/hr total
not reported
g6e.4xlargeAWS · eu-north-1
$3.19/GPU/hr
—
not reported
g6e.12xlargeAWS · ap-northeast-2
$3.23/GPU/hr
$12.90/hr total
not reported
g6e.2xlargeAWS · ap-northeast-1
$3.25/GPU/hr
—
not reported
g6e.2xlargeAWS · ap-northeast-3
$3.25/GPU/hr
—
not reported
g6e.12xlargeAWS · eu-central-1
$3.28/GPU/hr
$13.12/hr total
not reported
BM.GPU.L40S.4Oracle Cloud · global
$3.50/GPU/hr
$14.00/hr total
not reported
g6e.4xlargeAWS · ap-south-1
$3.61/GPU/hr
—
not reported
g6e.4xlargeAWS · ap-northeast-2
$3.69/GPU/hr
—
not reported
g6e.4xlargeAWS · eu-central-1
$3.76/GPU/hr
—
not reported
g6e.24xlargeAWS · us-east-1
$3.77/GPU/hr
$15.07/hr total
not reported
g6e.48xlargeAWS · us-east-1
$3.77/GPU/hr
$30.13/hr total
not reported
g6e.24xlargeAWS · us-east-2
$3.77/GPU/hr
$15.07/hr total
not reported
g6e.48xlargeAWS · us-east-2
$3.77/GPU/hr
$30.13/hr total
not reported
g6e.24xlargeAWS · us-west-2
$3.77/GPU/hr
$15.07/hr total
not reported
g6e.48xlargeAWS · us-west-2
$3.77/GPU/hr
$30.13/hr total
not reported
g6e.12xlargeAWS · ap-northeast-1
$3.80/GPU/hr
$15.22/hr total
not reported
g6e.12xlargeAWS · ap-northeast-3
$3.80/GPU/hr
$15.22/hr total
not reported
g6e.24xlargeAWS · eu-north-1
$4.00/GPU/hr
$15.98/hr total
not reported
g6e.48xlargeAWS · eu-north-1
$4.00/GPU/hr
$31.96/hr total
not reported
g6e.4xlargeAWS · ap-northeast-1
$4.36/GPU/hr
—
not reported
g6e.4xlargeAWS · ap-northeast-3
$4.36/GPU/hr
—
not reported
g6e.24xlargeAWS · ap-south-1
$4.52/GPU/hr
$18.09/hr total
not reported
g6e.48xlargeAWS · ap-south-1
$4.52/GPU/hr
$36.19/hr total
not reported
g6e.8xlargeAWS · us-east-1
$4.53/GPU/hr
—
not reported
Page 1 of 3

Agents can read the Markdown version, which lists every current instance.