Two Nvidia data center GPUs, one Ampere and one Grace Hopper: 80 GB HBM2e against 96 GB HBM3 and 2.1x the memory bandwidth on the GH200.
At a glance
Memory
- A100
- 80 GB
- GH200
- 96 GB
1.2x more on GH200
Memory bandwidth
- A100
- 1.94 TB/s
- GH200
- 4.00 TB/s
2.1x higher on GH200
Cheapest on-demand
- A100
- $0.89 /GPU/hr
- GH200
- $2.29 /GPU/hr
2.6x more expensive on GH200
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | A100 | GH200 | Difference |
|---|---|---|---|
| On-demand | $1.81 /GPU/hr from $0.89 | $2.52 /GPU/hr from $2.29 | 28% cheaper on A100 |
| Reserved | $1.60 /GPU/hr from $0.66 (24 mo) | No reserved listings | |
| Spot | $1.20 /GPU/hr from $0.67 | Sold out |
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. Verify before provisioning. More on how we price.
Price history
A100 vs GH200 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A100 | GH200 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 80 GB HBM2e 40 GB on SXM, PCIe | 96 GB HBM3 | 1.2x more on GH200 |
| Bandwidth | 1,935 GB/s 2,039 GB/s on SXM 1,555 GB/s on SXM, PCIe | 4,000 GB/s | 2.1x higher on GH200 |
| Bus width | 5,120-bit | ||
| Compute · dense | |||
| FP8 | Not supported | 1,979 TFLOPS | GH200 only |
| FP16 / BF16 | 312 TFLOPS | 989.5 TFLOPS | 3.2x higher on GH200 |
| INT8 | 624 TOPS | 1,979 TOPS | 3.2x higher on GH200 |
| FP32 | 19.5 TFLOPS | 67 TFLOPS | 3.4x higher on GH200 |
| FP64 | 9.7 TFLOPS | 34 TFLOPS | 3.5x higher on GH200 |
| Platform | |||
| Interconnect | NVLink 600 GB/s | NVLink 900 GB/s | 1.5x higher on GH200 |
| Released | Q2 2020 | Q2 2023 | 3 years newer on GH200 |
| Precision | |||
| Supported formats | FP16BF16TF32FP32FP64INT4INT8 | FP8FP16BF16TF32FP32FP64INT8 | |
Figures are for each model's headline configuration. Where a form factor carries a different number, it is named under the figure.
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: A100 datasheet and GH200 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A100 /GPU/hr | GH200 /GPU/hr |
|---|---|---|
Sesterce
|
On-Demand $1.49 | On-Demand $2.52 |
GPU.ai
|
On-Demand $0.89 | -- |
Thunder Compute
|
On-Demand $1.09 | -- |
Massed Compute
|
On-Demand $1.35 | -- |
Packet·ai
|
On-Demand $1.43 | -- |
Spheron
|
On-Demand $1.43 | -- |
Oblivus
|
On-Demand $1.47 | -- |
Geodd
|
On-Demand $1.49 | -- |
Lambda Labs
|
-- | On-Demand $2.29 |
CoreWeave
|
-- | On-Demand $6.50 |
Vultr
|
-- | Spot $1.99 |
43 providers list the A100 and 4 the GH200. Every listing is on the A100 page and the GH200 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia A100 was last updated on Sept. 7, 2026, and for Nvidia GH200 on Sept. 7, 2026.