AMD MI300X
vs
Nvidia H100
Two data center GPUs, one CDNA 3 and one Hopper: 192 GB HBM3 against 80 GB HBM3 and 58% more memory bandwidth on the MI300X.
At a glance
Updated 6 minutes ago- Memory
- 192 GB
- HBM3
- Memory bandwidth
- 5.30 TB/s
- From
- $2.59 /GPU/hr
- 10 providers
- Memory
- 80 GB
- HBM3 · 94 GB on NVL
- Memory bandwidth
- 3.35 TB/s
- 2,000 GB/s on PCIe · 3,938 GB/s on NVL
- From
- $1.50 /GPU/hr
- 53 providers
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | MI300X | H100 | Difference |
|---|---|---|---|
| On-demand | $2.59 | $1.50 | 42% cheaper on H100 |
| Spot | $1.60 | $0.17 | 9.3x more expensive on MI300X |
| Reserved | $1.61 | $1.53 | 5% cheaper on H100 |
| Median | $2.99 | $3.39 | 12% cheaper on MI300X |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
MI300X vs H100 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | MI300X | H100 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 192 GB HBM3 | 80 GB HBM3 94 GB on NVL | 2.4x more on MI300X |
| Bandwidth | 5,300 GB/s | 3,350 GB/s 2,000 GB/s on PCIe 3,938 GB/s on NVL | 1.6x higher on MI300X |
| Bus width | 8,192-bit | 5,120-bit | 1.6x wider on MI300X |
| Compute · dense | |||
| FP8 | 2,614.9 TFLOPS | 1,979 TFLOPS | 1.3x higher on MI300X |
| FP16 / BF16 | 1,307.4 TFLOPS | 989.5 TFLOPS | 1.3x higher on MI300X |
| INT8 | 2,614.9 TOPS | 1,979 TOPS | 1.3x higher on MI300X |
| FP32 | 163.4 TFLOPS | 67 TFLOPS | 2.4x higher on MI300X |
| FP64 | 81.7 TFLOPS | 34 TFLOPS | 2.4x higher on MI300X |
| Platform | |||
| Interconnect | Infinity Fabric 896 GB/s | NVLink 900 GB/s NVLink 600 GB/s on PCIe, NVL | Within 5% |
| Board power | 750 W | 700 W | 1.1x higher on MI300X |
| Released | Q4 2023 | Q3 2022 | 15 months newer on MI300X |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | 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: MI300X datasheet and H100 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | MI300X /GPU/hr | H100 /GPU/hr |
|---|---|---|
DigitalOcean
|
On-Demand $2.59 | -- |
Cyfuture AI
|
On-Demand $2.78 | -- |
Hot Aisle
|
On-Demand $2.99 | -- |
Oracle Cloud
|
On-Demand $6.00 | -- |
Azure
|
On-Demand $7.86 | -- |
Cirrascale
|
Reserved (12mo) $3.12 | -- |
Crusoe
|
Reservation on request | -- |
Runpod
|
Reservation on request | -- |
Lium
|
-- | On-Demand $1.50 |
HyperAI
|
-- | On-Demand $1.80 |
GPU.ai
|
-- | On-Demand $1.98 |
Koyeb
|
-- | On-Demand $2.50 |
Spheron
|
-- | On-Demand $2.65 |
Vast.ai
|
-- | On-Demand $2.67 |
Massed Compute
|
-- | On-Demand $2.73 |
Runcrate
|
-- | On-Demand $2.75 |
10 providers list the MI300X and 53 the H100. Every listing is on the MI300X page and the H100 page.
Similar GPUs
Other comparisons in the same class.
Our data for AMD MI300X was last updated on Sept. 7, 2026, and for Nvidia H100 on Sept. 7, 2026.