Supermicro’s application-optimized servers, featuring NVIDIA GPUs, will enable fine-tuning of pre-trained models and deploy AI inference solutions at the edge where data is generated to improve response times and decision making.

Super Micro Computer, Inc. (NASDAQ: SMCI) today announced that it has expanded its portfolio of AI solutions to support leveraging AI performance and capabilities in edge locations such as public spaces, retail stores, or industrial infrastructure.

“Supermicro has the broadest portfolio of Edge AI solutions and can support customers with pre-trained models tailored to their Edge environments,” said Charles Liang, President and CEO of Supermicro. He added, “Supermicro continues to provide industry-optimized solutions that enable enterprises to build competitiveness by processing AI data at the edge.”

Supermicro’s cutting-edge servers help bring needed information back to the edge without having to send the data back to the cloud for processing.

With NVIDIA AI Enterprise, customers can leverage performance-optimized and pre-trained Large Language Models (LLMs) directly at the edge, enabling real-time, accurate decision making close to the source of the data.

“Interest in leveraging AI at the edge is growing across industries including healthcare, retail, manufacturing and automotive,” said Kevin Connors, vice president of Worldwide OEM Accounts at NVIDIA. “The new Supermicro portfolio has passed the NVIDIA Certification System and is designed to deliver the highest level of accelerated computing infrastructure based on the NVIDIA AI platform.”

“It also comes with NVIDIA AI Enterprise software, making it easier to run edge AI workloads,” he said.

A prime example is Supermicro’s Hyper-Edge server, the SYS-221HE, which is optimized for edge training and inference and supports dual-socket CPUs in a shallow front I/O system.

The system can support up to three double-width NVIDIA Tensor Core GPUs, including NVIDIA H100, A10, L40S, A40, and A2 GPUs. The GPUs provide Supermicro Hyper-E with compute performance to handle AI workloads in edge environments where data is collected, analyzed, and stored.

“Eviden’s AI-driven retail solution, powered by Supermicro edge systems and NVIDIA technology, is revolutionizing how people navigate and interact with spaces like stores,” said Jacque Istok, CEO of StoreGenius. “It personalizes and enhances the shopping experience by delivering relevant information through 3D models of the store and conversational chatbots.”

He added that “by combining lifelike facial animation, advanced speech recognition and 3D modeling, we can make virtual shopping as realistic as visiting a physical store.”