NVIDIA L40 48GB GDDR6 PCIe Gen4 Professional GPU
Brand: NVIDIA | Category: GPUs
SKU: 900-2G133-0010-000 | Part #: 900-2G133-0010-000 | MPN: 900-2G133-0010-000
Contact for Pricing — Request a Quote
About the NVIDIA L40 48GB GDDR6 PCIe Gen4 Professional GPU
AI inference, real-time video processing, and professional visualization workloads benefit from the NVIDIA L40's Ada Lovelace architecture, which combines high-bandwidth memory with advanced tensor acceleration. This full-height, full-length, dual-slot passive-cooled GPU delivers 18,176 CUDA cores and 48 GB GDDR6 memory with ECC, making it a powerful addition to enterprise data centers requiring reliable, compute-intensive operations. For IT procurement teams and AI infrastructure specialists, the L40 (part number 900-2G133-0010-000) offers the memory capacity and bandwidth needed for large-scale batch processing without thermal management complexity.
The NVIDIA L40 features a 384-bit memory interface and 864 GB/s bandwidth, supporting demanding parallel workloads with FP32 performance of 90.5 TFLOPS. Its 4th Generation Tensor Cores deliver up to 362 TFLOPS in BFLOAT16 operations (724 TFLOPS with sparsity), while 3rd Generation RT Cores (142 cores total) accelerate ray tracing and AI-enhanced rendering. Eight NVENC video encode engines handle AV1, H.265, and H.264 simultaneously, and one NVDEC decode engine processes the same codecs plus VP9. The PCIe Gen4 x16 interface ensures low-latency data movement, and the 300 W maximum power envelope simplifies power supply planning in dense deployments.
Typical Enterprise Deployment Scenarios
- Large-language model inference and real-time AI recommendation engines requiring 48 GB unified memory
- Professional video transcoding, encoding, and streaming infrastructure with multi-codec support
- 3D rendering and digital content creation pipelines leveraging tensor and RT core acceleration
- Financial modeling, data analytics, and HPC simulations benefiting from high memory bandwidth and CUDA core density
- Graphics-intensive virtual workstation and CAD acceleration for distributed teams
To integrate the NVIDIA L40 into your infrastructure, contact Omnixon Global for a tailored quotation and technical consultation.
Technical Specifications
| Brand | NVIDIA |
| Category | GPUs |
| SKU | 900-2G133-0010-000 |
| Part Number | 900-2G133-0010-000 |
| Condition | New |
| Manufacturer Part Number | 900-2G133-0010-000 |
| GPU Architecture | Ada Lovelace |
| CUDA Cores | 18176 |
| GPU Memory | 48 GB GDDR6 with ECC |
| Memory Interface Width | 384-bit |
| Memory Bandwidth | 864 GB/s |
| FP32 Performance | 90.5 TFLOPS |
| TF32 Tensor Core Performance | 181 TFLOPS (sparsity: 362 TFLOPS) |
| BFLOAT16 Tensor Core Performance | 362 TFLOPS (sparsity: 724 TFLOPS) |
| INT8 Tensor Core Performance | 724 TOPS (sparsity: 1457 TOPS) |
| RT Core Generation | 3rd Generation (142 RT Cores) |
| Tensor Core Generation | 4th Generation (568 Tensor Cores) |
| Video Encode Engines | 8× NVENC (AV1, H.265, H.264) |
| Video Decode Engines | 1× NVDEC (AV1, H.265, H.264, VP9) |
| Form Factor | Full-height, Full-length (FHFL), Dual-slot, Passive Cooling |
| Interface | PCIe Gen4 x16 |
| Maximum Power Consumption | 300 W |
| Display Outputs | None (compute/professional GPU) |
| NVLink Support | Not supported |
| vGPU Software Support | Yes (NVIDIA AI Enterprise) |
| Operating System Support | Linux, Windows Server |
Frequently Asked Questions about NVIDIA L40 48GB GDDR6 PCIe Gen4 Professional GPU
What server platforms accept the NVIDIA L40 48GB GDDR6 PCIe Gen4 Professional GPU?
Reference servers include Dell PowerEdge XE9680 / XE9712, HPE Cray XD670, Lenovo ThinkSystem SR685a / SR675 V3, Supermicro AS-A21GE / SYS-821GE, Gigabyte G593 / G894, ASUS ESC. Share your target platform in the RFQ and we will confirm chassis-to-GPU compatibility and recommended NIC pairing.
How long is the lead time on AI GPUs?
Highly model-dependent. L40S / RTX-class: typically 3-6 weeks. H100/H200/B200 in SXM form factor: 12-16 weeks for whole-platform allocations. We quote genuine-channel ETAs only — no grey-market promises.
Do you supply matched networking (Quantum InfiniBand / Spectrum-X)?
Yes — Omnixon stocks the full NVIDIA networking lineup (Quantum-2 / Quantum-X InfiniBand, Spectrum-X Ethernet, ConnectX NICs, BlueField DPUs) so we can quote a complete training-cluster BOM, not just the GPUs.
Can you help with NVIDIA AI Enterprise licensing?
Yes. We hold genuine-channels for NVIDIA AI Enterprise software subscriptions. Add it to your RFQ and we quote node-aligned licensing along with the hardware.