A data center GPU against a consumer card, one Ada Lovelace and one Ampere: 48 GB GDDR6 against 8 GB GDDR6 and 93% more memory bandwidth on the L40.
At a glance
Memory
- L40
- 48 GB
- RTX 3070
- 8 GB
6x more on L40
Memory bandwidth
- L40
- 864 GB/s
- RTX 3070
- 448 GB/s
1.9x higher on L40
Cheapest on-demand
- L40
- $0.33 /GPU/hr
- RTX 3070
- $0.08 /GPU/hr
4.2x more expensive on L40
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | L40 | RTX 3070 | Difference |
|---|---|---|---|
| On-demand | $1.00 /GPU/hr from $0.33 | $0.10 /GPU/hr from $0.08 | 10x more expensive on L40 |
| Reserved | -- /GPU/hr from $0.70 | -- /GPU/hr from $0.10 (1 mo) | |
| Spot | $0.79 /GPU/hr from $0.77 | -- /GPU/hr from $0.04 |
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Medians are across providers, one vote each, so a large catalog does not outweigh a small one. No median for reserved on the L40 and reserved and spot on the RTX 3070: we only show one when at least 3 providers list the GPU on that billing type. Verify before provisioning. More on how we price.
Price history
L40 vs RTX 3070 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | L40 | RTX 3070 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 48 GB GDDR6 | 8 GB GDDR6 | 6x more on L40 |
| Bandwidth | 864 GB/s | 448 GB/s | 1.9x higher on L40 |
| Bus width | 384-bit | 256-bit | 1.5x wider on L40 |
| Compute · dense | |||
| FP8 | 362 TFLOPS | Not supported | L40 only |
| FP16 / BF16 | 181.1 TFLOPS | 40.6 TFLOPS | 4.5x higher on L40 |
| INT8 | 362 TOPS | 162.6 TOPS | 2.2x higher on L40 |
| FP32 | 90.5 TFLOPS | 20.3 TFLOPS | 4.5x higher on L40 |
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Board power | 300 W | 220 W | 1.4x higher on L40 |
| Released | Q4 2022 | Q4 2020 | 2 years newer on L40 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32INT4INT8 | FP16BF16TF32FP32INT4INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: L40 datasheet and RTX 3070 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | L40 /GPU/hr | RTX 3070 /GPU/hr |
|---|---|---|
Runpod
|
On-Demand $0.82 | On-Demand $0.13 |
Lium
|
On-Demand $0.33 | -- |
EmpirioLabs AI
|
On-Demand $0.70 | -- |
GPU.ai
|
On-Demand $0.78 | -- |
Thunder Compute
|
On-Demand $0.79 | -- |
Massed Compute
|
On-Demand $0.86 | -- |
Sesterce
|
On-Demand $0.97 | -- |
Runcrate
|
On-Demand $0.97 | -- |
Theta EdgeCloud
|
-- | On-Demand $0.08 |
Salad
|
-- | On-Demand $0.10 |
Vast.ai
|
-- | Reserved (1mo) $0.10 |
13 providers list the L40 and 4 the RTX 3070. Every listing is on the L40 page and the RTX 3070 page.
Similar GPUs
Other comparisons in the same class.
-
Nvidia A100
From $0.80 /GPU/hr Compare -
Nvidia L40S
From $0.38 /GPU/hr Compare -
Nvidia A40
From $0.49 /GPU/hr Compare -
Nvidia RTX 4070
From $0.15 /GPU/hr Compare -
Nvidia RTX 3060 Ti
From $0.08 /GPU/hr Compare
Our data for Nvidia L40 was last updated on Sept. 7, 2026, and for Nvidia RTX 3070 on Sept. 7, 2026.