4 GPU PCIe AI server, 2U
- GPUs
- 4 x RTX PRO 6000 Blackwell Server Edition
- CPU
- 2 x AMD EPYC 9005 or 2 x Intel Xeon 6
- Memory
- Up to 1.5 TB DDR5 RDIMM
- Chassis
- 2U, about 3.5 kW, air cooled
NVIDIA L40S
The Ada Lovelace universal data center GPU with 48 GB GDDR6 ECC, 1,466 TFLOPS sparse FP8 and 350 W, for inference, rendering and video in PCIe servers.
- Family
- Ada Lovelace
- Edition
- PCIe data center card, passive
- Memory
- 48 GB GDDR6 with ECC per GPU
- Memory bandwidth
- 864 GB/s per GPU
- Interconnect
- PCIe Gen4 x16 at 64 GB/s bidirectional, no NVLink
- Form factor
- Dual slot passive card, 4.4 x 10.5 in, full height full length
- TDP
- 350 W
- Cooling
- Passive
- Best for
- Universal inference, graphics and video servers where a proven 350 W PCIe GPU is wanted
- Availability
- Build to order
L40S is NVIDIA's universal data center GPU of the Ada Lovelace generation. It has 48 GB of GDDR6 with ECC at 864 GB/s, 18,176 CUDA cores, 568 fourth generation Tensor cores rated at 1,466 TFLOPS of sparse FP8, 142 third generation RT cores, and three NVENC and three NVDEC engines with AV1 support. It draws 350 W, sits in a dual slot passive form factor, and supports vGPU but not MIG or NVLink.
It remains a sensible choice for inference of small and mid sized models, for virtual desktops and cloud gaming, for rendering and for video transcoding, and it is widely available in validated PCIe servers. Where FP4 throughput or 96 GB of memory per card matters, RTX PRO 6000 Blackwell Server Edition is the successor. We supply L40S across the 1U, 2U, 4U and 5U inference server classes, from the single GPU edge server to the 10 GPU 5U server.
Systems that carry this GPU
4 GPU PCIe AI server, 4U
- GPUs
- 4 x RTX PRO 6000 Blackwell Server Edition
- CPU
- 2 x AMD EPYC 9005 or 2 x Intel Xeon 6
- Memory
- Up to 2 TB DDR5 RDIMM
- Chassis
- 4U, about 3.5 kW, air cooled
2 GPU 1U inference server
- GPUs
- 2 x NVIDIA L40S (48 GB)
- CPU
- 1 or 2 x AMD EPYC 9005, or 2 x Intel Xeon 6
- Memory
- Up to 768 GB DDR5 RDIMM
- Chassis
- 1U, about 1.2 kW, air cooled
2 GPU 2U inference server
- GPUs
- 2 x RTX PRO 6000 Blackwell Server Edition (96 GB)
- CPU
- 1 or 2 x AMD EPYC 9005, or 2 x Intel Xeon 6
- Memory
- Up to 1.5 TB DDR5 RDIMM
- Chassis
- 2U, about 2 kW, air cooled
Single GPU 1U edge inference server
- GPU
- 1 x NVIDIA L40S (48 GB)
- CPU
- 1 x AMD EPYC 9005 or 1 x Intel Xeon 6
- Memory
- Up to 384 GB DDR5 RDIMM
- Chassis
- 1U, about 0.8 kW, air cooled
10 GPU PCIe AI server, 5U
- GPUs
- 10 x RTX PRO 6000 Blackwell Server Edition (960 GB total)
- CPU
- 2 x AMD EPYC 9005 or 2 x Intel Xeon 6
- Memory
- Up to 2 TB DDR5 RDIMM
- Chassis
- 5U, about 7.5 kW, air cooled
Want this GPU in a different chassis?
We quote every class that carries it, from a single node to a pod. Tell us the workload and we propose the chassis.
- (858) 613-1818
- sales@computermodules.com
- Monday to Friday, 8am to 5pm Pacific