Two Nvidia data center GPUs, one Blackwell Ultra and one Hopper: up to 288 GB HBM3e against 80 GB HBM2e and 4x the memory bandwidth on the B300.
At a glance
Updated 6 minutes ago- Memory
- 288 GB
- HBM3e
- Memory bandwidth
- 8.00 TB/s
- From
- $6.50 /GPU/hr
- 30 providers
- Memory
- 80 GB
- HBM2e
- Memory bandwidth
- 2.00 TB/s
- From
- --
Price comparison
Cheapest listed price per GPU per hour by billing type, and the median across providers.
| Billing | B300 | H800 | Difference |
|---|---|---|---|
| On-demand | $6.50 | -- | |
| Spot | $2.99 | No spot listings | |
| Reserved | $4.25 | No reserved listings | |
| Median | $7.87 | -- |
Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.
Price history
B300 vs H800 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | B300 | H800 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 288 GB HBM3e | 80 GB HBM2e | 3.6x more on B300 |
| Bandwidth | 8,000 GB/s | 2,000 GB/s | 4x higher on B300 |
| Bus width | 8,192-bit | 5,120-bit | 1.6x wider on B300 |
| Compute · dense | |||
| FP4 | 14,000 TFLOPS | Not supported | B300 only |
| FP8 | 4,500 TFLOPS | 1,513 TFLOPS | 3x higher on B300 |
| FP16 / BF16 | 2,250 TFLOPS | 756.5 TFLOPS | 3x higher on B300 |
| INT8 | 153.5 TOPS | 1,513 TOPS | 9.9x higher on H800 |
| FP32 | 75 TFLOPS | 51 TFLOPS | 1.5x higher on B300 |
| FP64 | 1.2 TFLOPS | ||
| Platform | |||
| Interconnect | NVLink 1,800 GB/s | NVLink 400 GB/s | 4.5x higher on B300 |
| Board power | 1,100 W | 350 W | 3.1x higher on B300 |
| Released | Q1 2025 | Q1 2023 | 2 years newer on B300 |
| Precision | |||
| Supported formats | FP4FP6FP8FP16BF16TF32FP32FP64INT8 | FP8FP16BF16TF32FP32FP64INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: B300 datasheet .
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | B300 /GPU/hr | H800 /GPU/hr |
|---|---|---|
GPU.ai
|
On-Demand $7.89 | -- |
Runpod
|
On-Demand $7.89 | -- |
Lyceum
|
On-Demand $7.99 | -- |
GPUaaS
|
Custom on request | -- |
Packet·ai
|
Custom on request | -- |
Impossible Cloud
|
Reserved (60mo) $4.32 | -- |
Canopy Wave
|
Custom on request | -- |
Bentaus
|
On-Demand $6.50 | -- |
30 providers list the B300 and 0 the H800. Every listing is on the B300 page and the H800 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia B300 was last updated on Sept. 7, 2026, and for Nvidia H800 on Sept. 7, 2026.