Super Micro Computer, Inc. (SMCI), a global leader in total IT solutions for AI/ML, Cloud, Storage, and 5G/Edge, has added plug-and-play AI infrastructure solutions for the NVIDIA Omniverse platform to its Supercluster portfolio.

The new Supercluster solution delivers high-performance, generative AI-based 3D workflows at enterprise scale. It also features the latest Supermicro NVIDIA OVX systems, enabling enterprises to easily scale as their workloads grow.

“Supermicro has led the industry in developing GPU-optimized products for 3D graphics and application acceleration, and now we’re developing products for AI,” said Charles Liang, President and CEO of Supermicro. “With the advent of AI, enterprises are looking for compute infrastructure that combines all capabilities in a single package. Supermicro superclusters scale up to 256 NVIDIA L40S PCIe GPUs and feature fully interconnected 4U PCIe GPU NVIDIA-certified systems for NVIDIA Omniverse, supporting high performance across the Omniverse platform, including generative AI integration.” “Supermicro’s development of a supercluster for Omniverse is not simply a product offering, but a bridgehead toward the future of application development and innovation,” he explained.

Supermicro has expanded its portfolio of application-optimized AI rack solutions with the launch of Superclusters for NVIDIA Omniverse. A wide range of professionals rely on compute-intensive 3D workflows for a variety of applications, from product design to industrial digital twins. Generative AI is ushering in a new era of applications while enhancing existing 3D workflows. Superclusters for NVIDIA Omniverse simplify the deployment of scalable infrastructure to meet the multi-workload demands of 3D and AI.

Supermicro NVIDIA OVX systems serve as the fundamental building blocks of cluster computing performance. Each system node features up to eight of the latest NVIDIA PCIe GPUs, delivering ultimate 3D performance and exceptional generative AI performance with Tensor Cores and Transformer Engines. The system is powered by four 2700W Titanium Level PSUs in a well-ventilated chassis, ensuring reliability even in high-use scenarios. Up to four NVIDIA BlueField-3 SuperNICs or four NVIDIA ConnectX-7 NICs per system provide 400Gb/s networking speeds along with high scalability and security.

Supermicro’s 4U PCIe GPU systems have undergone a rigorous validation process for performance, reliability, scalability, and security and are NVIDIA certified to utilize NVIDIA Omniverse. Enterprises can maximize performance for a variety of workloads, including generative AI technologies leveraging the NVIDIA Omniverse development platform and the globally deployed OpenUSD ecosystem and Omniverse cloud APIs.

The NVIDIA Omniverse Supercluster is a fully interconnected infrastructure solution that gives developers, artists, engineers, and more access to the highest level of GPU computing on demand, offering seamless access to virtual GPUs or bare metal access to entire system nodes. Additionally, the 400Gb/s high-performance network fabric supports the NVIDIA Spectrum-X Ethernet platform. This allows companies developing custom large-scale language models (LLMs) to leverage a pool of GPU memory combined across system nodes, which is essential for training large-scale AI models.

Supermicro’s validated rack solutions range from 4 GPUs to 256 GPU scalable units, with additional scalability to meet the needs of any size enterprise. Customers receive a plug-and-play rack that is fully tested and validated at the L12 level, ready to go from day one.