A data center GPU against a consumer card, one Grace Blackwell and one Ampere: 186 GB HBM3e against 24 GB GDDR6X and 8.5x the memory bandwidth on the GB200.
At a glance
Updated 6 minutes ago- Memory
- 186 GB
- HBM3e
- Memory bandwidth
- 8.00 TB/s
- From
- $10.50 /GPU/hr
- 8 providers
- Memory
- 24 GB
- GDDR6X
- Memory bandwidth
- 936 GB/s
- From
- $0.10 /GPU/hr
- 10 providers
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | GB200 | RTX 3090 | Difference |
|---|---|---|---|
| On-demand | $10.50 | $0.10 | 105x more expensive on GB200 |
| Spot | $27.04 | $0.07 | 397x more expensive on GB200 |
| Reserved | $10.58 | $0.06 | 176x more expensive on GB200 |
| Median | $16.00 | $0.25 | 64x more expensive on GB200 |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
GB200 vs RTX 3090 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | GB200 | RTX 3090 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 186 GB HBM3e | 24 GB GDDR6X | 7.8x more on GB200 |
| Bandwidth | 8,000 GB/s | 936 GB/s | 8.5x higher on GB200 |
| Bus width | 384-bit | ||
| Compute · dense | |||
| FP4 | 10,000 TFLOPS | Not supported | GB200 only |
| FP8 | 5,000 TFLOPS | Not supported | GB200 only |
| FP16 / BF16 | 2,500 TFLOPS | 71.2 TFLOPS | 35x higher on GB200 |
| INT8 | 5,000 TOPS | 284.7 TOPS | 18x higher on GB200 |
| FP32 | 80 TFLOPS | 35.6 TFLOPS | 2.2x higher on GB200 |
| FP64 | 40 TFLOPS | Not supported | GB200 only |
| Platform | |||
| Interconnect | NVLink 1,800 GB/s | NVLink 112 GB/s | 16x higher on GB200 |
| Board power | 1,200 W | 350 W | 3.4x higher on GB200 |
| Released | Q1 2024 | Q3 2020 | 4 years newer on GB200 |
| Precision | |||
| Supported formats | FP4FP6FP8FP16BF16TF32FP32FP64INT8 | FP16BF16TF32FP32INT4INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: GB200 datasheet and RTX 3090 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | GB200 /GPU/hr | RTX 3090 /GPU/hr |
|---|---|---|
CoreWeave
|
On-Demand $10.50 | -- |
Oracle Cloud
|
On-Demand $16.00 | -- |
Azure
|
On-Demand $27.04 | -- |
AWS
|
Reservation $10.58 | -- |
CUDO
|
Custom on request | -- |
Canopy Wave
|
Custom on request | -- |
Nebius
|
Custom on request | -- |
Together AI
|
Custom on request | -- |
HyperAI
|
-- | On-Demand $0.10 |
Vast.ai
|
-- | On-Demand $0.12 |
Lium
|
-- | On-Demand $0.13 |
Theta EdgeCloud
|
-- | On-Demand $0.13 |
GPU.ai
|
-- | On-Demand $0.21 |
Salad
|
-- | On-Demand $0.25 |
EmpirioLabs AI
|
-- | On-Demand $0.45 |
SwissGPU
|
-- | On-Demand $0.31 |
8 providers list the GB200 and 10 the RTX 3090. Every listing is on the GB200 page and the RTX 3090 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia GB200 was last updated on Sept. 7, 2026, and for Nvidia RTX 3090 on Sept. 7, 2026.