B300
NVIDIA · Blackwell Ultra
288 GBHBM3e
8 TB/sBandwidth
4.5 PFFP8 dense
$4,899 /GPU/mo$6,552 · −25%
NVIDIA · Ada Lovelace · PCIe 4.0 · up to 8 per server
Rent the NVIDIA L40S by the month.
The NVIDIA L40S is our entry GPU: 48 GB of GDDR6 and 0.73 PF of dense FP8 on Ada Lovelace, from the lowest monthly price in the catalog. It is a reliable choice for small-model inference, diffusion and 3D rendering.
48 GBGDDR6 per GPU
864 GB/sMemory bandwidth
0.73 PFFP8 dense
PCIe 4.0 x16Interconnect
Why the L40S
What the L40S is best at
- 48 GB GDDR6. Room for 7B–13B models in BF16 or 30B at 8-bit.
- Lowest monthly price. The cheapest dedicated data-center GPU we rent.
- Graphics + AI. RT cores and NVENC for rendering and video alongside inference.
- Small and mid-size model inference
- Stable Diffusion and video generation
- 3D rendering, VFX and visualization
PCIe 4.0, up to 8 cards per server. Ideal for scale-out inference behind a load balancer.
Pricing
L40S monthly price by server size
One flat price per GPU, the same from 1 to 8 GPUs. The market median is the public on-demand list price of specialist GPU clouds × 730 hours.
| Server | CoreGPU / month | Market median / month | You save / month | You save / year | ≈ per GPU-hour |
|---|---|---|---|---|---|
| 1× L40S | $849 | $297 | $3,564 | $1.16 | |
| 2× L40S | $1,698 | $594 | $7,128 | $1.16 | |
| 4× L40S | $3,396 | $1,188 | $14,256 | $1.16 | |
| 8× L40S | $6,792 | $2,376 | $28,512 | $1.16 |
Prices in USD, excluding taxes. Market median checked Nov 4, 2026 across specialist GPU clouds; hyperscalers excluded. All GPU prices
Compare
L40S vs RTX PRO 6000
L40S costs less per month; RTX PRO 6000 doubles the memory and adds FP4 for larger models.
| L40S | RTX PRO 6000 | |
|---|---|---|
| Architecture | Ada Lovelace | Blackwell |
| GPU memory | 48 GB GDDR6 | 96 GB GDDR7 |
| Memory bandwidth | 864 GB/s | 1.6 TB/s |
| FP8 dense | 0.73 PF | 1 PF |
| FP4 dense | n/a | 2 PF |
| Interconnect | PCIe 4.0 x16 | PCIe 5.0 x16 |
| Power (TDP) | 350 W | 600 W |
| Price per GPU / month | $849 | $1,199 |
| Price per GB of memory | $17.69 | $12.49 |
| Price per FP8 PFLOPS | $1,163 | $1,199 |
| Availability | 38 available | 52 available |
Specifications
NVIDIA L40S technical specs
Per-GPU figures from the vendor datasheet. Tensor throughput is dense (without sparsity).
| Model | NVIDIA L40S |
|---|---|
| Architecture | Ada Lovelace |
| Form factor | PCIe 4.0 · up to 8 per server |
| GPU memory | 48 GB GDDR6 |
| Memory bandwidth | 864 GB/s |
| FP8 Tensor (dense) | 0.73 PF |
| FP4 Tensor (dense) | Not supported |
| Interconnect | PCIe 4.0 x16 |
| Power (TDP) | 350 W |
| GPUs per server | 1 to 8 |
| Availability | 38 available |
FAQ
L40S rental: common questions
Another question? Our engineers answer within one business day.
Contact usHow much does it cost to rent an NVIDIA L40S per month?
$849 per GPU per month at CoreGPU, about $1.16 per GPU-hour, versus a market median of $1,146 ($1.57/hour × 730 h). A full 8-GPU server is $6,792 per month, taxes excluded.
Is the L40S available right now?
Yes: 38 available in the current allocation of 128 GPUs. In-stock servers are delivered with root SSH access in under 60 seconds after payment.
How many L40S GPUs can I rent?
From 1 to 8 GPUs per server, at the same price per GPU. These are PCIe cards, so they scale per GPU rather than over NVLink. Need more? Rent several servers or ask for a private allocation.
L40S or RTX PRO 6000?
L40S costs less per month; RTX PRO 6000 doubles the memory and adds FP4 for larger models. See the full L40S vs RTX PRO 6000 comparison.
What do I get with a L40S server?
Dedicated GPUs (never shared or sliced), root SSH access, your choice of a clean OS or a ready image with CUDA, PyTorch and vLLM, local NVMe scratch, and month-to-month billing with no setup fee.
Alternatives
Other GPUs in stock now
Deploy your L40S server in under 60 seconds
Pick 1 to 8 GPUs in the configurator above. Month to month, cancel anytime.