900-2G133-0080-000
- Description
- L40S 48GB Accelerator Card
- Chipset Manufacturer
- Nvidia
- Model Number
- L40S
Quantity
390 AvailableLet us do the sourcing. Even if it's not available now, our team can track it down and tailor a solution just for you.
- 48GB GDDR6 memory with ECC support
- Ada Lovelace GPU architecture
- 18176 CUDA cores, 568 Tensor cores, 142 RT cores
- Dual-slot PCIe 4.0 x16 form factor
- 864GB/s memory bandwidth
- Passive cooling for server airflow
- vGPU software support
The NVIDIA 900-2G133-0080-000 is an L40S accelerator card featuring 48GB of GDDR6 memory with ECC support and the Ada Lovelace GPU architecture. It provides 18176 CUDA cores, 568 Tensor cores, and 142 RT cores across a dual-slot PCIe 4.0 x16 card. With 864GB/s of memory bandwidth and passive cooling, it is designed for high-density server chassis.
The L40S is used for AI training and inference, generative AI, and graphics and rendering workloads in datacenter servers. Its ECC memory and passive-cooled dual-slot design suit rack servers running virtualized GPU workloads. It is deployed by cloud providers and enterprises building AI and visual-computing infrastructure.
Fusion’s in-house quality hubs and Prosemi testing facility are fully certified to meet critical industry standards for electronic component inspection and testing.
Quantity
390 AvailableLet us do the sourcing. Even if it's not available now, our team can track it down and tailor a solution just for you.