Two Nvidia data center GPUs, one Blackwell Ultra and one Ada Lovelace: up to 288 GB HBM3e against 48 GB GDDR6, 9.3x the memory bandwidth on the B300, and NVLink 1,800 GB/s on the B300 where the L40 is PCIe only.
At a glance
Memory
- B300
- 288 GB
- L40
- 48 GB
6x more on B300
Memory bandwidth
- B300
- 8,000 GB/s
- L40
- 864 GB/s
9.3x higher on B300
Cheapest on-demand
- B300
- $6.50 /GPU/hr
- L40
- $0.78 /GPU/hr
8.3x more expensive on B300
Price comparison
Median price per GPU per hour by billing type, with the cheapest listing under it.
| Billing | B300 | L40 | Difference |
|---|---|---|---|
| On-demand | $7.89 /GPU/hr from $6.50 | $1.05 /GPU/hr from $0.78 | 7.5x more expensive on B300 |
| Reserved | $5.71 /GPU/hr from $4.25 | -- /GPU/hr from $0.70 | |
| Spot | $5.85 /GPU/hr from $4.30 | $0.79 /GPU/hr from $0.77 | 7.4x more expensive on B300 |
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. No median for reserved on the L40: we only show one when at least 3 providers list the GPU on that billing type. Verify before provisioning. More on how we price.
Price history
B300 vs L40 price history
Weekly median on-demand price per GPU per hour
Specs comparison
| Spec | B300 | L40 | Difference |
|---|---|---|---|
| Memory | |||
| Capacity | 288 GB HBM3e | 48 GB GDDR6 | 6x more on B300 |
| Bandwidth | 8,000 GB/s | 864 GB/s | 9.3x higher on B300 |
| Bus width | 8,192-bit | 384-bit | 21x wider on B300 |
| Compute · dense | |||
| FP4 | 14,000 TFLOPS | Not supported | B300 only |
| FP8 | 4,500 TFLOPS | 362 TFLOPS | 12x higher on B300 |
| FP16 / BF16 | 2,250 TFLOPS | 181.1 TFLOPS | 12x higher on B300 |
| INT8 | 153.5 TOPS | 362 TOPS | 2.4x higher on L40 |
| FP32 | 75 TFLOPS | 90.5 TFLOPS | 1.2x higher on L40 |
| FP64 | 1.2 TFLOPS | Not supported | B300 only |
| Platform | |||
| Interconnect | NVLink 1,800 GB/s | PCIe only | NVLink on B300 only |
| Board power | 1,100 W | 300 W | 3.7x higher on B300 |
| Released | Q1 2025 | Q4 2022 | 2 years newer on B300 |
| Precision | |||
| Supported formats | FP4FP6FP8FP16BF16TF32FP32FP64INT8 | FP8FP16BF16TF32FP32INT4INT8 | |
Compute: Vendor peak figures, dense and per GPU. Not measured throughput.
Source: B300 datasheet and L40 datasheet.
Where to rent them
Providers listing these GPUs, those with both first, so you can switch without moving clouds.
| Provider | B300 /GPU/hr | L40 /GPU/hr |
|---|---|---|
GPU.ai
|
On-Demand $7.89 | On-Demand $0.78 |
EmpirioLabs AI
|
On-Demand $16.00 | On-Demand $1.50 |
Spheron
|
Spot $5.85 | On-Demand $1.06 |
Runpod
|
On-Demand $7.89 | -- |
Lium
|
On-Demand $8.00 | -- |
Vast.ai
|
On-Demand $8.75 | -- |
GPUaaS
|
Custom on request | -- |
Packet·ai
|
Custom on request | -- |
Thunder Compute
|
-- | On-Demand $0.79 |
Massed Compute
|
-- | On-Demand $0.86 |
Sesterce
|
-- | On-Demand $0.97 |
Runcrate
|
-- | On-Demand $0.97 |
Oblivus
|
-- | On-Demand $1.05 |
33 providers list the B300 and 12 the L40. Every listing is on the B300 page and the L40 page.
Similar GPUs
Other comparisons in the same class.
Our data for Nvidia B300 was last updated on Sept. 7, 2026, and for Nvidia L40 on Sept. 7, 2026.