A2 vs RTX 5090
The cheapest A2 ($0.046) costs 79% less per hour than the cheapest RTX 5090 ($0.22).
Per unit of performance, the RTX 5090 is 63% cheaper: $0.24 versus $0.64 per A100-equivalent hour.
The RTX 5090 has 32 GB of memory against 16 GB on the A2.
| Measure | NVIDIA A2 16GB | NVIDIA GeForce RTX 5090 |
|---|---|---|
| Lowest per GPU-hour | $0.046 | $0.22 |
| Median per GPU-hour | $0.23 | $0.87 |
| On-demand from | $0.046 | $0.34 |
| Spot from | — | $0.25 |
| Reserved from | $0.23 | $0.22 |
| Lowest per A100-equivalent hour | $0.64 | $0.24 |
| Live offers | 9 | 405 |
| Providers | 7 | 84 |
| Memory | 16 GB GDDR6 | 32 GB GDDR7 |
| Memory bandwidth | 200 GB/s | 1,792 GB/s |
| FP16 / BF16 | 18 TFLOPS | 209.5 TFLOPS |
| FP8 | — | 419 TFLOPS |
| Performance vs A100 80GB | 0.07× | 0.92× |
| Architecture | Ampere | Blackwell |
| Power | 60 W | 575 W |