NVIDIA
A10 24GB
Tier 1-9
- Memory
- 48 GB GDDR6 ECC
- Memory bandwidth
- 864 GB/s
- TDP
- 350 W
- Interface
- PCIe Gen4 x16, dual-slot
NVIDIA L40S 48GB
P/N900-2G133-0080-000
In stockStock tier 100+
Request a quote now
Price and availability confirmed within one business day
- Manufacturer channel supply
- Ships from Dubai / Hong Kong
- Commercial invoice, packing list, warranty docs
01Overview
L40S is a universal Ada Lovelace data center GPU that combines AI inference and fine-tuning performance with full graphics and video capabilities. It is chosen when one server must handle generative AI, rendering, virtual workstations and video pipelines, and when a 350 W air-cooled PCIe card fits an existing fleet better than Hopper-class accelerators.
Built for
- LLM and generative AI inference for models up to ~48 GB of weights per card
- Fine-tuning of small and mid-size models with FP8 Tensor Cores
- 3D rendering, Omniverse and real-time ray tracing on the server
- Virtual workstations and vGPU deployments (NVIDIA RTX vWS)
- Video transcoding and streaming with AV1 encode/decode
02Specifications
Technical specifications
Compute
- Architecture
- NVIDIA Ada Lovelace
- CUDA cores
- 18,176
- Tensor Cores (4th gen)
- 568
- RT Cores (3rd gen)
- 142
- FP32
- 91.6 TFLOPS
- TF32 Tensor Core
- 183 / 366 TFLOPSwith sparsity
- FP16 Tensor Core
- 362 / 733 TFLOPSwith sparsity
- FP8 / INT8 Tensor Core
- 733 / 1,466 TFLOPSwith sparsity
- RT Core performance
- 212 TFLOPS
Memory
- Capacity
- 48 GB GDDR6 with ECC
- Bandwidth
- 864 GB/s
- MIG
- Not supported
Interface & form factor
- Host interface
- PCIe Gen4 x16, 64 GB/s bidirectional
- Form factor
- 4.4" (H) × 10.5" (L), dual-slot
- Display outputs
- 4× DisplayPort 1.4a
- NVLink
- Not supported
Power & cooling
- Max power consumption
- 350 W
- Power connector
- 16-pin
- Thermal design
- Passive — requires server chassis airflow
- NEBS
- Level 3 ready
Software & features
- Media engines
- 3× NVENC, 3× NVDEC (AV1 encode/decode)
- vGPU
- Supported
- Secure boot
- Yes, with root of trust
03Requirements
Installation
- Server with a free dual-slot PCIe Gen4 (or Gen5) x16 FHFL slot and a 16-pin auxiliary power feed rated for 350 W
- Passive cooling — chassis airflow must be designed for a 350 W card; not for desktops without forced airflow
- For vGPU or NVIDIA AI Enterprise deployments a separate software license is required
In the box
Manufacturer retail or bulk packaging as supplied through the channel; contents confirmed on the order.
04Documents
Manufacturer
Warranty: Manufacturer warranty, details on quote
Supplied with every shipment
- Commercial invoice and packing list
- Manufacturer warranty registration documents
- Certificate of origin on request
- Serial numbers listed on the delivery note
05Related
More in GPUs & accelerators
NVIDIA
A100 80GB SXM
Tier 1-9
Tier 10+
Tier 1-9