AMD MI325X
vs
Nvidia RTX 5090
The MI325X has 8x the memory and 3.3x the bandwidth. The RTX 5090 is typically cheaper.
At a glance
Memory
8x more on MI325X
- MI325X
- 256 GB
- RTX 5090
- 32 GB
Memory bandwidth
3.3x higher on MI325X
- MI325X
- 6.00 TB/s
- RTX 5090
- 1.79 TB/s
Cheapest on-demand
84% cheaper on RTX 5090
- MI325X
- $2.25 /GPU/hr
- RTX 5090
- $0.35 /GPU/hr
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | MI325X | RTX 5090 | Difference |
|---|---|---|---|
| On-demand | $3.06 /GPU/hr from $2.25 | $0.69 /GPU/hr from $0.35 | 78% cheaper on RTX 5090 |
| Reserved | $1.95 /GPU/hr from $1.48 (12 mo) | $0.59 /GPU/hr from $0.21 (1 mo) | 70% cheaper on RTX 5090 |
| Spot | Sold out | $0.48 /GPU/hr from $0.25 |
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
MI325X vs RTX 5090 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | MI325X | RTX 5090 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 256 GB HBM3e | 32 GB GDDR7 | 8x more on MI325X |
| Bandwidth | 6,000 GB/s | 1,792 GB/s | 3.3x higher on MI325X |
| Bus width | 8,192-bit | 512-bit | 16x wider on MI325X |
| Compute · dense | |||
| FP4 | Not supported | 1,676 TFLOPS | RTX 5090 only |
| FP8 | 2,614.9 TFLOPS | 419 TFLOPS | 6.2x higher on MI325X |
| FP16 / BF16 | 1,307.4 TFLOPS | 209.5 TFLOPS | 6.2x higher on MI325X |
| INT8 | 2,614.9 TOPS | 838 TOPS | 3.1x higher on MI325X |
| FP32 | 163.4 TFLOPS | 104.8 TFLOPS | 56% higher on MI325X |
| FP64 | 81.7 TFLOPS | Not supported | MI325X only |
| Platform | |||
| Interconnect | Infinity Fabric 896 GB/s | PCIe only | Infinity Fabric on MI325X only |
| Board power | 1,000 W | 575 W | 74% higher on MI325X |
| Released | Q4 2024 | Q1 2025 | |
| Precision | |||
| Supported formats | FP8FP16BF16TF32FP32FP64INT8 | FP4FP8FP16BF16TF32FP32INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: MI325X datasheet and RTX 5090 datasheet.
Where to rent them
Providers that rent these GPUs, with each one's cheapest rate. No provider we track offers both right now.
| Provider | MI325X | RTX 5090 |
|---|---|---|
Bentaus
USA
|
$2.25 /GPU/hr On-demand | -- |
Cyfuture AI
India
|
$3.01 /GPU/hr On-demand | -- |
DigitalOcean
USA
|
$3.80 /GPU/hr On-demand | -- |
TensorWave
USA
|
$2.25 /GPU/hr Custom OAM | -- |
Vultr
USA
|
$2.00 /GPU/hr Spot | -- |
HyperAI
Hong Kong
|
-- | $0.35 /GPU/hr On-demand |
GPU.ai
UAE
|
-- | $0.41 /GPU/hr On-demand |
Vast.ai
USA
|
-- | $0.43 /GPU/hr On-demand |
GPUhub
Singapore
|
-- | $0.46 /GPU/hr On-demand |
Theta EdgeCloud
Distributed
|
-- | $0.65 /GPU/hr On-demand |
Nova Cloud
Canada
|
-- | $0.66 /GPU/hr On-demand PCIe |
Runpod
USA
|
-- | $0.69 /GPU/hr On-demand |
Sesterce
France
|
-- | $0.71 /GPU/hr On-demand |
5 providers list the MI325X and 23 the RTX 5090. Every listing is on the MI325X page and the RTX 5090 page.
Similar GPUs
Other comparisons in the same class.
-
Nvidia H100
80 GB HBM3 · 44% less bandwidth than the MI325X
From $1.73 /GPU/hr Compare
-
Nvidia B200
180 GB HBM3e · 28% more bandwidth than the MI325X
From $3.75 /GPU/hr Compare
-
Nvidia H200
141 GB HBM3e · 20% less bandwidth than the MI325X
From $2.09 /GPU/hr Compare
-
Nvidia RTX PRO 6000
96 GB GDDR7 · workstation card
From $0.65 /GPU/hr Compare
-
Nvidia RTX 4090
24 GB GDDR6X · 44% less bandwidth than the RTX 5090
From $0.33 /GPU/hr Compare
Our data for AMD MI325X was last updated on Sept. 14, 2026, and for Nvidia RTX 5090 on Sept. 14, 2026.