Two Nvidia data center GPUs, one Hopper and one Volta: 141 GB HBM3e against 32 GB HBM2 and 5.3x the memory bandwidth on the H200.
At a glance
Memory
- H200
- 141 GB
- V100
- 32 GB
4.4x more on H200
Memory bandwidth
- H200
- 4,800 GB/s
- V100
- 900 GB/s
5.3x higher on H200
Cheapest on-demand
- H200
- $2.09 /GPU/hr
- V100
- $0.14 /GPU/hr
15x more expensive on H200
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | H200 | V100 | Difference |
|---|---|---|---|
| On-demand | $4.40 /GPU/hr from $2.09 | $1.57 /GPU/hr from $0.14 | 2.8x more expensive on H200 |
| Reserved | $3.65 /GPU/hr from $2.29 (1 mo) | $1.51 /GPU/hr from $0.13 (24 mo) | 2.4x more expensive on H200 |
| Spot | $3.01 /GPU/hr from $1.60 | $0.43 /GPU/hr from $0.09 | 7x more expensive on H200 |
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
H200 vs V100 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | H200 | V100 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 141 GB HBM3e | 32 GB HBM2 | 4.4x more on H200 |
| Bandwidth | 4,800 GB/s | 900 GB/s | 5.3x higher on H200 |
| Bus width | 4,096-bit | ||
| Compute · dense | |||
| FP8 | 1,979 TFLOPS | Not supported | H200 only |
| FP16 / BF16 | 989.5 TFLOPS | 125 TFLOPS | 7.9x higher on H200 |
| INT8 | 1,979 TOPS | ||
| FP32 | 67 TFLOPS | 15.7 TFLOPS | 4.3x higher on H200 |
| FP64 | 34 TFLOPS | 7.8 TFLOPS | 4.4x higher on H200 |
| Platform | |||
| Interconnect | NVLink 900 GB/s | NVLink 300 GB/s | 3x higher on H200 |
| Board power | 700 W | 300 W | 2.3x higher on H200 |
| Released | Q4 2023 | Q2 2017 | 6 years newer on H200 |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP16FP32FP64INT8 | |
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: H200 datasheet and V100 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | H200 /GPU/hr | V100 /GPU/hr |
|---|---|---|
Verda
|
On-Demand $4.00 | On-Demand $0.17 |
Runcrate
|
On-Demand $4.40 | On-Demand $2.57 |
Lium
|
On-Demand $2.90 | -- |
Koyeb
|
On-Demand $3.00 | -- |
GPU.ai
|
On-Demand $3.29 | -- |
Massed Compute
|
On-Demand $3.62 | -- |
Vast.ai
|
On-Demand $3.97 | -- |
Lyceum
|
On-Demand $4.29 | -- |
HyperAI
|
-- | On-Demand $0.14 |
Sesterce
|
-- | On-Demand $2.57 |
Geodd
|
-- | On-Demand $0.16 |
TensorDock
|
-- | On-Demand $0.28 |
Cyfuture AI
|
-- | On-Demand $0.55 |
OVH
|
-- | On-Demand $0.77 |
46 providers list the H200 and 19 the V100. Every listing is on the H200 page and the V100 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia H200 was last updated on Sept. 7, 2026, and for Nvidia V100 on Sept. 7, 2026.