Two Nvidia data center GPUs, one Grace Hopper and one Ada Lovelace: 96 GB HBM3 against 48 GB GDDR6, 4.6x the memory bandwidth on the GH200, and NVLink 900 GB/s on the GH200 where the L40 is PCIe only.
At a glance
Memory
- GH200
- 96 GB
- L40
- 48 GB
2x more on GH200
Memory bandwidth
- GH200
- 4,000 GB/s
- L40
- 864 GB/s
4.6x higher on GH200
Cheapest on-demand
- GH200
- $2.29 /GPU/hr
- L40
- $0.78 /GPU/hr
2.9x more expensive on GH200
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | GH200 | L40 | Difference |
|---|---|---|---|
| On-demand | $2.52 /GPU/hr from $2.29 | $1.05 /GPU/hr from $0.78 | 2.4x more expensive on GH200 |
| Reserved | No reserved listings | -- /GPU/hr from $0.70 | |
| Spot | Sold out | $0.79 /GPU/hr from $0.77 |
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: 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
GH200 vs L40 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | GH200 | L40 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 96 GB HBM3 | 48 GB GDDR6 | 2x more on GH200 |
| Bandwidth | 4,000 GB/s | 864 GB/s | 4.6x higher on GH200 |
| Bus width | 384-bit | ||
| Compute · dense | |||
| FP8 | 1,979 TFLOPS | 362 TFLOPS | 5.5x higher on GH200 |
| FP16 / BF16 | 989.5 TFLOPS | 181.1 TFLOPS | 5.5x higher on GH200 |
| INT8 | 1,979 TOPS | 362 TOPS | 5.5x higher on GH200 |
| FP32 | 67 TFLOPS | 90.5 TFLOPS | 1.4x higher on L40 |
| FP64 | 34 TFLOPS | Not supported | GH200 only |
| Platform | |||
| Interconnect | NVLink 900 GB/s | PCIe only | NVLink on GH200 only |
| Released | Q2 2023 | Q4 2022 | 7 months newer on GH200 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP8FP16BF16TF32FP32INT4INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: GH200 datasheet and L40 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | GH200 /GPU/hr | L40 /GPU/hr |
|---|---|---|
Sesterce
|
On-Demand $2.52 | On-Demand $0.97 |
Lambda Labs
|
On-Demand $2.29 | -- |
CoreWeave
|
On-Demand $6.50 | -- |
Vultr
|
Spot $1.99 | -- |
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 |
EmpirioLabs AI
|
-- | On-Demand $1.50 |
4 providers list the GH200 and 12 the L40. Every listing is on the GH200 page and the L40 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia GH200 was last updated on Sept. 7, 2026, and for Nvidia L40 on Sept. 7, 2026.