GPU comparison
NVIDIA H100 vs NVIDIA RTX 5090.
Specs, performance and live cloud availability, side by side.
The short answer
Choose NVIDIA H100 if you need 80GB+ on a single GPU (48GB more than NVIDIA RTX 5090).
Choose NVIDIA H100 if you need FP16 tensor throughput (372% faster than NVIDIA RTX 5090).
Compare another pair
vs
Hopper
NVIDIA H100
from$4.29/hr
Available now · 4 regions- Memory
- 80 GB HBM3
- FP16 tensor
- up to 989 TFLOPS
- Bandwidth
- up to 3.4 TB/s
Blackwell
NVIDIA RTX 5090
Not on demand yet- Memory
- 32 GB GDDR7
- FP16 tensor
- 209.5 TFLOPS
- Bandwidth
- 1.8 TB/s
At a glance
Spec by spec. The longer bar is the better number.
Full specs
The complete spec sheet.
| Spec | NVIDIA H100 | NVIDIA RTX 5090 |
|---|---|---|
| Architecture | Hopper | Blackwell |
| Memory | 80 GB | 32 GB |
| Memory type | HBM3 | GDDR7 |
| Memory bandwidth | up to 3.4 TB/s | 1.8 TB/s |
| FP32 | up to 67 TFLOPS | 104.8 TFLOPS |
| FP16 tensor | up to 989 TFLOPS | 209.5 TFLOPS |
| INT8 | up to 2400 TOPS | 838 TOPS |
| TDP | up to 700 W | 575 W |
| GPUs per node | 8 | 2 |
| Hourly price (live) | from $4.29/hr | Not on demand yet |
Questions
H100 or RTX 5090. What buyers ask first.
Other comparisons