Two Nvidia data center GPUs, one Ampere and one Grace Hopper: 48 GB GDDR6 against 96 GB HBM3, 5.7x the memory bandwidth on the GH200, and NVLink 900 GB/s on the GH200 where the A40 is PCIe only.
At a glance
Updated 6 minutes ago- Memory
- 48 GB
- GDDR6
- Memory bandwidth
- 696 GB/s
- From
- $0.49 /GPU/hr
- 6 providers
- Memory
- 96 GB
- HBM3
- Memory bandwidth
- 4.00 TB/s
- From
- $6.50 /GPU/hr
- 4 providers
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | A40 | GH200 | Difference |
|---|---|---|---|
| On-demand | $0.49 | $6.50 | 13x more expensive on GH200 |
| Reserved | $0.61 | No reserved listings | |
| Median | $0.77 | $6.50 | 8.5x more expensive on GH200 |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
A40 vs GH200 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A40 | GH200 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 48 GB GDDR6 | 96 GB HBM3 | 2x more on GH200 |
| Bandwidth | 696 GB/s | 4,000 GB/s | 5.7x higher on GH200 |
| Bus width | 384-bit | ||
| Compute · dense | |||
| FP8 | Not supported | 1,979 TFLOPS | GH200 only |
| FP16 / BF16 | 149.7 TFLOPS | 989.5 TFLOPS | 6.6x higher on GH200 |
| INT8 | 299.3 TOPS | 1,979 TOPS | 6.6x higher on GH200 |
| FP32 | 37.4 TFLOPS | 67 TFLOPS | 1.8x higher on GH200 |
| FP64 | Not supported | 34 TFLOPS | GH200 only |
| Platform | |||
| Interconnect | PCIe only | NVLink 900 GB/s | NVLink on GH200 only |
| Released | Q4 2020 | Q2 2023 | 3 years newer on GH200 |
| Precision | |||
| Supported formats | FP16BF16TF32FP32INT4INT8 | FP8FP16BF16TF32FP32FP64INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: A40 datasheet and GH200 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A40 /GPU/hr | GH200 /GPU/hr |
|---|---|---|
Sesterce
|
On-Demand $1.21 | On-Demand $2.52 |
Runpod
|
On-Demand $0.49 | -- |
Exoscale
|
On-Demand $1.05 | -- |
Database Mart
|
Reserved (24mo) $0.61 | -- |
Utho
|
Reserved (36mo) $0.65 | -- |
EmpirioLabs AI
|
On-Demand $1.49 | -- |
CoreWeave
|
-- | On-Demand $6.50 |
Lambda Labs
|
-- | On-Demand $2.29 |
Vultr
|
-- | Spot $1.99 |
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia A40 was last updated on Sept. 7, 2026, and for Nvidia GH200 on Sept. 7, 2026.