NVIDIA L40 48 GB OEM - the same Ada Lovelace chip for server integrations
The OEM variant (tray, factory packaging) is intended for system integrators and batch server assembly, where a retail box with accessories is not needed. The accelerator itself is identical to the retail version: Ada Lovelace architecture, 18,176 CUDA cores, 568 4th-generation Tensor cores (FP8, BF16, TF32, structural sparsity) and 142 3rd-generation RT cores.
On board are 48 GB of GDDR6 with ECC, a 384-bit bus and 864 GB/s of bandwidth - a memory reserve for LLM training and inference, simulations, 3D rendering and NVIDIA vGPU profiles (vApps, vPC, RTX vWS) that share the GPU between several users. Three NVENC and three NVDEC units provide hardware encoding and decoding of AV1, H.265 and H.264 for streaming and video work.
- Architecture: NVIDIA Ada Lovelace, 18,176 CUDA cores
- 48 GB GDDR6 ECC, 384-bit, 864 GB/s
- 568 4th-gen Tensor cores, 142 3rd-gen RT cores
- PCIe Gen4 x16, 16-pin power connector, 300 W TGP
- Dual-slot, passive cooling, 4 x DisplayPort 1.4a
- Secure Boot with Root of Trust, NEBS Level 3
Passive dual-slot cooling and a 300 W TGP are designed for dense 24x7 deployment in NVIDIA-Certified Systems. Supplied in tray (OEM) format - the official distributor warranty is retained. Check batch availability and shipping dates by phone at +380 93,594 00 77.
Not sure this card fits your system?
We will check compatibility with your platform, calculate the power draw for your configuration and tell you whether the memory is enough for your models.
The model you need is out of stock? ETE is listed among official PNY partners, so we bring professional NVIDIA cards to order with the manufacturer warranty.
Message us on Telegram or leave a request.
