A data center GPU against a consumer card, one Ampere and one Ada Lovelace: 80 GB HBM2e against 24 GB GDDR6X, 92% more memory bandwidth on the A100, and NVLink 600 GB/s on the A100 where the RTX 4090 is PCIe only.
At a glance
Memory
- A100
- 80 GB
- RTX 4090
- 24 GB
3.3x more on A100
Memory bandwidth
- A100
- 1.94 TB/s
- RTX 4090
- 1.01 TB/s
1.9x higher on A100
Cheapest on-demand
- A100
- $0.33 /GPU/hr
- RTX 4090
- $0.27 /GPU/hr
20% cheaper on RTX 4090
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | A100 | RTX 4090 | Difference |
|---|---|---|---|
| On-demand | $1.80 /GPU/hr from $0.33 | $0.44 /GPU/hr from $0.27 | 4.1x more expensive on A100 |
| Reserved | $1.62 /GPU/hr from $0.46 (1 mo) | $0.49 /GPU/hr from $0.27 (12 mo) | 3.3x more expensive on A100 |
| Spot | $1.14 /GPU/hr from $0.14 | $0.17 /GPU/hr from $0.15 | 6.7x more expensive on A100 |
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
A100 vs RTX 4090 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | A100 | RTX 4090 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 80 GB HBM2e 40 GB on SXM, PCIe | 24 GB GDDR6X | 3.3x more on A100 |
| Bandwidth | 1,935 GB/s 2,039 GB/s on SXM 1,555 GB/s on SXM, PCIe | 1,010 GB/s | 1.9x higher on A100 |
| Bus width | 5,120-bit | 384-bit | 13x wider on A100 |
| Compute · dense | |||
| FP8 | Not supported | 330.3 TFLOPS | RTX 4090 only |
| FP16 / BF16 | 312 TFLOPS | 165.2 TFLOPS | 1.9x higher on A100 |
| INT8 | 624 TOPS | 660.6 TOPS | 1.1x higher on RTX 4090 |
| FP32 | 19.5 TFLOPS | 82.6 TFLOPS | 4.2x higher on RTX 4090 |
| FP64 | 9.7 TFLOPS | Not supported | A100 only |
| Platform | |||
| Interconnect | NVLink 600 GB/s | PCIe only | NVLink on A100 only |
| Board power | 300 W | 450 W | 1.5x higher on RTX 4090 |
| Released | Q2 2020 | Q4 2022 | 2 years newer on RTX 4090 |
| Precision | |||
| Supported formats | FP16BF16TF32FP32FP64INT4INT8 | FP8FP16BF16TF32FP32INT4INT8 | |
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: A100 datasheet and RTX 4090 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | A100 /GPU/hr | RTX 4090 /GPU/hr |
|---|---|---|
Vast.ai
|
On-Demand $0.33 | On-Demand $0.27 |
GPU.ai
|
On-Demand $0.41 | On-Demand $0.40 |
Lium
|
On-Demand $1.23 | On-Demand $0.30 |
Packet·ai
|
On-Demand $1.43 | On-Demand $0.39 |
Runpod
|
On-Demand $1.00 | -- |
Thunder Compute
|
On-Demand $1.09 | -- |
Massed Compute
|
On-Demand $1.35 | -- |
Spheron
|
On-Demand $1.43 | -- |
Novita
|
-- | On-Demand $0.33 |
Salad
|
-- | On-Demand $0.33 |
Theta EdgeCloud
|
-- | On-Demand $0.38 |
Sesterce
|
-- | On-Demand $0.54 |
44 providers list the A100 and 20 the RTX 4090. Every listing is on the A100 page and the RTX 4090 page.
Similar GPUs
Other comparisons in the same class.
-
Nvidia H100
80 GB HBM3 · 73% more bandwidth than the A100
From $1.30 /GPU/hr Compare -
Nvidia RTX 5090
32 GB GDDR7 · PCIe only
From $0.35 /GPU/hr Compare
-
AMD MI300X
192 GB HBM3 · 2.7x the bandwidth of the A100
From $2.39 /GPU/hr Compare -
Nvidia L40S
48 GB GDDR6 · 14% less bandwidth than the RTX 4090
From $0.38 /GPU/hr Compare
Our data for Nvidia A100 was last updated on Sept. 14, 2026, and for Nvidia RTX 4090 on Sept. 14, 2026.