Rent the RTX PRO 6000 if…
you need more than 48 GB — 70B at FP8, larger scenes, several resident models — or Blackwell features such as FP4, or Windows.
Deploy an RTX PRO 6000 · $749/mo
NVIDIA RTX PRO 6000 Blackwell96 GB GDDR7 ECC · Blackwell
vs
NVIDIA L40S48 GB GDDR6 · Ada Lovelace
RTX PRO 6000 vs L40S: is it time to upgrade?
Workstation-class Blackwell against the Ada data center all-rounder: 96 GB at about 1.6 TB/s with FP4, or 48 GB at 864 GB/s with FP8 — for 1.7× the monthly price.
- RTX PRO 6000, dedicated$749per month · $1.03/h effective
- L40S, dedicated$429per month · $0.59/h effective
- Price ratio1.7×the RTX PRO 6000 costs 1.7× the L40S per month
The short answer
Which one should you rent?
The RTX PRO 6000 Blackwell is the natural step up from the L40S: twice the memory, nearly twice the bandwidth and a newer generation of tensor and RT cores, for 1.7× the price. If your workload fits in 48 GB, the L40S remains the cheaper always-on GPU. If it does not — 70B-class models at FP8, big scenes, several models side by side — the RTX PRO 6000 holds it on one card.
- Memory. 96 GB against 48 GB, both with ECC.
- Bandwidth. About 1.6 TB/s of GDDR7 against 864 GB/s of GDDR6.
- Generation. Fifth-generation tensor cores with FP4 against fourth-generation with FP8; fourth- against third-generation RT cores.
- Price per GB. $7.80 per GB a month on the RTX PRO 6000 against $8.94 on the L40S.
Rent the L40S if…
your models fit in 48 GB and you want the lowest monthly price for always-on inference, generation or rendering.
Deploy an L40S · $429/moSpecifications
RTX PRO 6000 vs L40S: the numbers side by side
NVIDIA's published figures for each part, tensor throughput without sparsity. The last column is how the RTX PRO 6000 compares, as a multiple of the L40S.
| Specification | NVIDIA RTX PRO 6000 Blackwell | NVIDIA L40S | RTX PRO 6000 vs L40S |
|---|---|---|---|
| Architecture | Blackwell · 2025 | Ada Lovelace · 2023 | — |
| GPU memory | 96 GB GDDR7 ECC | 48 GB GDDR6 | 2× |
| Memory bandwidth | 1.6 TB/s | 864 GB/s | 1.9× |
| FP16 / BF16 tensor (dense) | ~500 TFLOPS | 362 TFLOPS | 1.4× |
| FP8 tensor (dense) | ~1,000 TFLOPS | 733 TFLOPS | 1.4× |
| FP4 tensor (dense) | ~2,000 TFLOPS | Not supported | Only the RTX PRO 6000 |
| Maximum power | Up to 600 W | 350 W | 1.7× |
| Form factor | PCIe Gen5 x16 | PCIe Gen4 x16 | — |
| GPU-to-GPU link | PCIe only | PCIe only | — |
| Multi-Instance GPU (MIG) | Up to 4 instances | No | — |
| ECC memory | Yes | Yes | — |
| Video engines | 4 NVENC · 4 NVDEC (AV1) | 3 NVENC · 3 NVDEC (AV1) | — |
| CUDA compute capability | 12.0 | 8.9 | — |
| Dedicated host here | 16 vCPU · 128 GB RAM · 2 TB NVMe | 12 vCPU · 96 GB RAM · 1 TB NVMe | — |
| Price per month here | $749 | $429 | 1.7× |
Sources: NVIDIA datasheets and product pages for each GPU. Tensor figures are dense (NVIDIA often prints them doubled “with sparsity”). Prices are this site's flat monthly rates for a dedicated server with the host listed, in USD excluding VAT.
Value
What each dollar buys
Price per month divided into the two resources most jobs run out of first — memory and memory bandwidth — and what the same GPU costs kept on for a month at the market-median hourly rate.
Per GB of GPU memoryRTX PRO 6000 $7.80 · L40S $8.94per month, dedicated host included
Per TB/s of bandwidthRTX PRO 6000 $468 · L40S $497per month — what decode-heavy inference pays for
Effective hourly rateRTX PRO 6000 $1.03/h · L40S $0.59/hflat monthly price over 730 hours
Same GPU on demand, 24/7RTX PRO 6000 $1,606 · L40S $1,124market-median hourly rate × 730 h
Model sizes
What fits on one RTX PRO 6000 and one L40S
The largest model whose weights fit on a single GPU, keeping 10% of the memory for the runtime, activations and a small KV cache. Long contexts and large batches need more — size them with the GPU memory calculator.
| Weights precision | NVIDIA RTX PRO 6000 Blackwell | NVIDIA L40S | Typical use |
|---|---|---|---|
| 16-bit (BF16 / FP16) | up to ~43B parameters | up to ~22B parameters | Full-precision inference; the base weights of a LoRA fine-tune |
| 8-bit (FP8 / INT8) | up to ~86B parameters | up to ~43B parameters | FP8 serving with vLLM or TensorRT-LLM on GPUs that support it; INT8 weights elsewhere |
| 4-bit (GPTQ, AWQ, Q4) | up to ~160B parameters | up to ~79B parameters | Quantised inference on a budget; QLoRA fine-tuning |
Parameters in billions, weights only: 2 bytes per parameter at 16-bit, 1 at FP8 or INT8, about 0.55 at 4-bit (GPTQ, AWQ, GGUF Q4). Mixture-of-experts models need memory for all their experts, not just the active ones.
Prices
RTX PRO 6000 and L40S price per month, by term
Per dedicated GPU server, in USD excluding VAT — CPU cores, RAM, NVMe, 20 TB of transfer and 24/7 support included.
| Term | NVIDIA RTX PRO 6000 Blackwell | NVIDIA L40S | Difference per month |
|---|---|---|---|
| Monthly | $749 | $429 | $320 |
| 3 months · −5% | $712 | $408 | $304 |
| 6 months · −10% | $674 | $386 | $288 |
| 12 months · −15% | $637 | $365 | $272 |
| Dedicated host | 16 vCPU · 128 GB RAM · 2 TB NVMe · 10 Gbps | 12 vCPU · 96 GB RAM · 1 TB NVMe · 10 Gbps | included |
RTX PRO 6000 vs L40S: questions answered
Is the RTX PRO 6000 the successor of the L40S?
In practice: it fills the same universal AI-and-graphics role one generation later, with twice the memory, nearly twice the bandwidth and FP4.
Can I run Windows on these GPUs?
On the RTX PRO 6000, yes: Windows Server 2022 or 2025 with the RTX driver and RDP is available at checkout for $29 a month. The L40S runs Linux images here.
Which is better for image generation?
Both run SDXL and FLUX comfortably. The RTX PRO 6000 is faster and keeps more models resident at once; the L40S costs less per month for a single pipeline.
Are the RTX PRO 6000 and the L40S in stock?
6 of 24 RTX PRO 6000 units were free at the last stock update; 17 of 40 L40S units were free at the last stock update (18 September 2026). Both are provisioned automatically, in under 10 minutes after payment.
Can I rent the RTX PRO 6000 or the L40S without KYC?
Yes. Both are rented with an email address and paid from a prepaid balance in Bitcoin, Ether or USDT — $749 and $429 a month, dedicated, with no identity check and no card.
Both are in stock, by the month
Dedicated, single-tenant and live in under 10 minutes: the RTX PRO 6000 at $749 or the L40S at $429 a month, paid in BTC, ETH or USDT with no KYC.