A data center GPU against a consumer card, one Ampere and one Blackwell: 40 GB HBM2 against 32 GB GDDR7, 15% more memory bandwidth on the RTX 5090, and NVLink 400 GB/s on the A800 where the RTX 5090 is PCIe only.
At a glance
Updated 6 minutes ago- Memory
- 40 GB
- HBM2
- Memory bandwidth
- 1.55 TB/s
- From
- $1.18 /GPU/hr
- 1 provider
- Memory
- 32 GB
- GDDR7
- Memory bandwidth
- 1.79 TB/s
- From
- $0.35 /GPU/hr
- 18 providers
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | A800 | RTX 5090 | Difference |
|---|---|---|---|
| On-demand | $1.18 | $0.35 | 3.4x more expensive on A800 |
| Spot | No spot listings | $0.25 | |
| Reserved | No reserved listings | $0.21 | |
| Median | $1.18 | $0.63 | 47% cheaper on RTX 5090 |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
A800 vs RTX 5090 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A800 | RTX 5090 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 40 GB HBM2 | 32 GB GDDR7 | 1.2x more on A800 |
| Bandwidth | 1,555 GB/s | 1,792 GB/s | 1.2x higher on RTX 5090 |
| Bus width | 5,120-bit | 512-bit | 10x wider on A800 |
| Compute · dense | |||
| FP4 | Not supported | 1,676 TFLOPS | RTX 5090 only |
| FP8 | Not supported | 419 TFLOPS | RTX 5090 only |
| FP16 / BF16 | 312 TFLOPS | 209.5 TFLOPS | 1.5x higher on A800 |
| INT8 | 623.5 TOPS | 838 TOPS | 1.3x higher on RTX 5090 |
| FP32 | 19.5 TFLOPS | 104.8 TFLOPS | 5.4x higher on RTX 5090 |
| FP64 | 9.7 TFLOPS | Not supported | A800 only |
| Platform | |||
| Interconnect | NVLink 400 GB/s | PCIe only | NVLink on A800 only |
| Board power | 240 W | 575 W | 2.4x higher on RTX 5090 |
| Released | Q4 2022 | Q1 2025 | 2 years newer on RTX 5090 |
| Precision | |||
| Supported formats | FP16BF16TF32FP32FP64INT4INT8 | FP4FP8FP16BF16TF32FP32INT8 | |
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: A800 datasheet and RTX 5090 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A800 /GPU/hr | RTX 5090 /GPU/hr |
|---|---|---|
GPUhub
|
On-Demand $1.18 | On-Demand $0.46 |
HyperAI
|
-- | On-Demand $0.35 |
Salad
|
-- | On-Demand $0.45 |
Nova Cloud
|
-- | On-Demand $0.63 |
Vast.ai
|
-- | On-Demand $0.73 |
GPU.ai
|
-- | On-Demand $53.34 |
GPUaaS
|
-- | Custom on request |
SwissGPU
|
-- | On-Demand $0.61 |
1 provider list the A800 and 18 the RTX 5090. Every listing is on the A800 page and the RTX 5090 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia A800 was last updated on Sept. 7, 2026, and for Nvidia RTX 5090 on Sept. 7, 2026.