Supermicro L40S GPU Add-in Board 100-300000070

The Supermicro L40S GPU Add-in Board (part number 100-300000070) is a high-density, full-height full-length PCIe add-in card built around the NVIDIA Ada Lovelace architecture. Featuring 48 GB of GDDR6 ECC memory across a 384-bit memory bus, the L40S delivers exceptional throughput for mixed inference and training workloads, with 91.6 TFLOPS of FP32 performance and support for NVIDIA's third-generation Tensor Cores. The card operates in a single-slot thermal design leveraging active cooling and is optimized for rack-dense enterprise server configurations, including Supermicro's own validated multi-GPU platforms.

The L40S is positioned as a universal data-center GPU, combining professional visualization capabilities with high-performance AI compute on a single card. It supports PCIe Gen 4 x16 connectivity and NVIDIA's full Ada Lovelace feature set including fourth-generation NVENC and NVDEC media engines, enabling simultaneous AI inference and real-time media transcoding without resource contention. The board implements ECC across its full memory capacity, making it suitable for mission-critical enterprise deployments where data integrity under sustained compute load is mandatory.

Designed for large-scale deployment in enterprise data centers across AI inferencing, high-performance computing, and professional visualization, the Supermicro 100-300000070 integrates within standard data-center power and cooling envelopes at a 350 W TDP. Its broad software ecosystem compatibility spans CUDA, TensorRT, NVIDIA AI Enterprise, and major virtualization stacks including vGPU, enabling flexible workload consolidation across bare-metal, virtualized, and containerized environments.

Ideal for

• Large language model (LLM) and generative AI inference serving at scale, leveraging 48 GB ECC GDDR6 memory to host multi-billion-parameter models on a single card
• High-throughput video transcoding and AI-accelerated media processing using fourth-generation NVENC/NVDEC engines in broadcast and streaming infrastructure
• Enterprise GPU virtualization with NVIDIA vGPU software, enabling multiple concurrent virtual workstation or inference instances on a single physical card
• Scientific and engineering HPC simulation workloads requiring high FP32 and FP64 throughput with memory error correction for computational integrity
• Professional 3D visualization and CAD/CAM rendering in virtualized or bare-metal workstation environments for design and manufacturing sectors
• AI-accelerated data analytics and machine learning model training pipelines in on-premises data centers where PCIe Gen 4 bandwidth and large VRAM are critical

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.