AMD MI325X
vs
Nvidia H20
Two data center GPUs, one CDNA 3 and one Hopper: 256 GB HBM3e against 96 GB HBM3, 50% more memory bandwidth on the MI325X, and Infinity Fabric 896 GB/s on the MI325X where the H20 is PCIe only.
At a glance
Updated 6 minutes ago- Memory
- 256 GB
- HBM3e
- Memory bandwidth
- 6.00 TB/s
- From
- $2.25 /GPU/hr
- 5 providers
- Memory
- 96 GB
- HBM3
- Memory bandwidth
- 4.00 TB/s
- From
- --
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | MI325X | H20 | Difference |
|---|---|---|---|
| On-demand | $2.25 | -- | |
| Reserved | $1.48 | No reserved listings | |
| Median | $3.07 | -- |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
MI325X vs H20 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | MI325X | H20 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 256 GB HBM3e | 96 GB HBM3 | 2.7x more on MI325X |
| Bandwidth | 6,000 GB/s | 4,000 GB/s | 1.5x higher on MI325X |
| Bus width | 8,192-bit | ||
| Platform | |||
| Interconnect | Infinity Fabric 896 GB/s | PCIe only | Infinity Fabric on MI325X only |
| Board power | 1,000 W | 400 W | 2.5x higher on MI325X |
| Released | Q4 2024 | Q1 2024 | 9 months newer on MI325X |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP8FP16BF16TF32FP32FP64INT8 | |
Source: MI325X datasheet .
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | MI325X /GPU/hr | H20 /GPU/hr |
|---|---|---|
Bentaus
|
On-Demand $2.25 | -- |
Cyfuture AI
|
On-Demand $3.02 | -- |
DigitalOcean
|
On-Demand $3.80 | -- |
TensorWave
|
Custom $2.25 | -- |
Vultr
|
Spot $2.00 | -- |
Similar GPUs
Other comparisons in the same class.
Our data for AMD MI325X was last updated on Sept. 7, 2026, and for Nvidia H20 on Sept. 7, 2026.