| Flavour Name | GPUs | vCPUs | RAM | Hourly | Monthly | 6 Monthly 5% Off | 12 Monthly 10% Off | |
|---|---|---|---|---|---|---|---|---|
| N.L40S.64 | 1x | 16 | 64 | --- | ₹83,000 | ₹473,100 | ₹896,400 | |
| N.L40S.128 | 1x | 16 | 128 | --- | ₹100,876 | ₹574,991 | ₹1,089,457 | |
| N.L40S.192 | 2x | 24 | 192 | --- | ₹183,110 | ₹1,043,726 | ₹1,977,586 | |
| N.L40S.256 | 2x | 32 | 256 | --- | ₹201,708 | ₹1,149,734 | ₹2,178,442 | |
| N.L40S.512 | 4x | 64 | 512 | --- | ₹403,372 | ₹2,299,218 | ₹4,356,414 |
Build an hourly and monthly estimate for your exact setup across compute, GPU, storage, and networking. Configure your resources and get a real-time total with a simple cost breakdown.
Pricing CalculatorAceCloud’s NVIDIA L40S 48GB cloud GPU pricing starts at ₹60,000/month. This plan includes 1x NVIDIA L40S GPU, 16 vCPUs, and 64 GB RAM.
AceCloud lists 1x, 2x, and 4x NVIDIA L40S 48GB Linux cloud GPU plans. The available configurations range from 1x GPU with 16 vCPUs / 64 GB RAM to 4x GPUs with 64 vCPUs / 512 GB RAM.
NVIDIA L40S comes with 48GB GDDR6 memory with ECC and 864 GB/s memory bandwidth. NVIDIA also lists it with 18,176 CUDA cores, 568 fourth-generation Tensor Cores, and 142 third-generation RT Cores.
Yes. NVIDIA L40S is designed for generative AI and LLM training and inference. Its 48GB GPU memory, FP8 support, fourth-generation Tensor Cores, and Ada Lovelace architecture make it useful for many AI workloads that need more VRAM than smaller GPUs.
NVIDIA L40S is a newer Ada Lovelace data center GPU optimized for AI, graphics, rendering, and simulation workloads. NVIDIA lists L40S with 48GB GDDR6 ECC memory, 91.6 FP32 TFLOPS, and 1,466 FP8 Tensor TFLOPS with sparsity, making it a strong option for mixed AI and visualization workloads.
Choose a 1x L40S plan for workloads that fit within one 48GB GPU, such as inference, rendering, model testing, and development. Choose a 2x or 4x L40S plan when your workload can scale across multiple GPUs or needs more vCPUs, RAM, and parallel GPU capacity.
