Catalogue image. Confirm the exact supplied variant against the part number.
NVIDIA L40S GPU is a universal data-centre accelerator for AI inference/training, rendering, virtual workstations and graphics-heavy enterprise workloads.
Price covers: L40S 48GB GDDR6 PCIe 4.0, passive — 900-2G133-0080-000. New card; other configurations quoted separately.
Enquire about the named variant or request a quotation for another configuration. No payment is taken.
NVIDIA L40S GPU is a universal data-centre accelerator for AI inference/training, rendering, virtual workstations and graphics-heavy enterprise workloads.
Confirm the server model, GPU form factor, riser and supported configuration. Accelerator compatibility depends on the complete platform.
This is the exact catalogue listing for NVIDIA-L40S. Detailed electrical, physical and performance specifications must be confirmed using the manufacturer documentation for this part number.
Confirm the server model, GPU form factor, riser and supported configuration. Accelerator compatibility depends on the complete platform.
Passive and active cooling designs require different environments. Check the server thermal matrix, GPU power cabling and system power budget.
Identify the operating system, driver, framework and any virtualisation or AI software licence requirements as part of the specification.
Catalogue prices were recorded on 7 October 2026. Contact us for current availability, a final price and delivery arrangements for your exact part number and quantity.
Add products to your bag and review your selection, then send an order enquiry. We will confirm price, availability and next steps by email. Sending an enquiry does not place an order or reserve stock.
Price covers: H100 80GB PCIe, passive — 900-21010-0000-000. New card; other configurations quoted separately.
Price covers: H200 NVL 141GB PCIe — 900-21010-0040-000. New card; other configurations quoted separately.
Price covers: A100 80GB PCIe, passive — Supermicro GPU-NVA100-80-NC. New card; other configurations quoted separately.