Two Nvidia data center GPUs, one Hopper and one Ada Lovelace: 96 GB HBM3 against 48 GB GDDR6 and 4.6x the memory bandwidth on the H20.
At a glance
Updated 7 minutes ago- Memory
- 96 GB
- HBM3
- Memory bandwidth
- 4.00 TB/s
- From
- --
- Memory
- 48 GB
- GDDR6
- Memory bandwidth
- 864 GB/s
- From
- $0.78 /GPU/hr
- 12 providers
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | H20 | L40 | Difference |
|---|---|---|---|
| On-demand | -- | $0.78 | |
| Spot | No spot listings | $0.78 | |
| Reserved | No reserved listings | $0.70 | |
| Median | -- | $1.05 |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
H20 vs L40 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | H20 | L40 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 96 GB HBM3 | 48 GB GDDR6 | 2x more on H20 |
| Bandwidth | 4,000 GB/s | 864 GB/s | 4.6x higher on H20 |
| Bus width | 384-bit | ||
| Platform | |||
| Interconnect | PCIe only | PCIe only | |
| Board power | 400 W | 300 W | 1.3x higher on H20 |
| Released | Q1 2024 | Q4 2022 | 15 months newer on H20 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP8FP16BF16TF32FP32INT4INT8 | |
Source: L40 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | H20 /GPU/hr | L40 /GPU/hr |
|---|---|---|
GPU.ai
|
-- | On-Demand $0.78 |
Thunder Compute
|
-- | On-Demand $0.79 |
Massed Compute
|
-- | On-Demand $0.86 |
Runcrate
|
-- | On-Demand $0.97 |
Oblivus
|
-- | On-Demand $1.05 |
Spheron
|
-- | On-Demand $1.06 |
Sesterce
|
-- | On-Demand $1.10 |
EmpirioLabs AI
|
-- | On-Demand $1.50 |
0 providers list the H20 and 12 the L40. Every listing is on the H20 page and the L40 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia H20 was last updated on Sept. 7, 2026, and for Nvidia L40 on Sept. 7, 2026.