This local AI supercomputer packs data-center-class performance and flexible deployment options to make it the premier solution for neocloud operators and AI natives looking for an on-ramp to AI Factories.

Supermicro returns to Asia’s premier ICT trade show, COMPUTEX 2026 this week, where AI and in particular agentic AI are set to take center stage. First unveiled at GTC Spring in March 2026 and making its Asia debut at COMPUTEX, Supermicro’s Super AI Station is a tower-sized local AI supercomputer powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip that can be deployed deskside or mounted in a 5U rack slot for quick and easy upscaling. The Super AI Station features native support for the NVIDIA AI Enterprise software suite that streamlines agentic AI development and deployment. The result is a powerful, flexible local AI development platform perfect for AI natives building agentic AI applications as well as enterprises and neocloud operators looking to deploy cost-effective supporting infrastructure.

Data-Center-Level Performance with Flexible, Scalable Deployment Options

Supermicro has reimagined what an AI development platform can be, drawing on years of R&D expertise to develop a local AI supercomputer that addresses fundamental challenges faced by smaller-scale AI developers and AI natives, while also providing enhanced flexibility for infrastructure operators:

• Uncompromising performance for AI development:

  A coherent memory architecture comprising 252GB of HBM3e GPU memory and 496GB of LPDDR5X system memory, creating a shared memory pool of 748GB to support AI training datasets and up to 1 trillion-parameter frontier models. With 20 PFLOPS of uncompromising FP4 AI performance, the Super AI Station delivers true data-center-class performance in a deskside, flexible form factor.

• Flexible and scalable deployment options:

  The Super AI Station can be deployed either as a deskside tower in offices, laboratories, and other space-constrained environments, or mounted in a rack for clustering and upscaling. Based on the same NVIDIA GB300 architecture that powers AI factories around the world, AI natives can develop and deploy their AI agents and workloads to AI Factories without the need for re-architecting. Integrated dual port 400Gbps high-speed networking with NVIDIA ConnectX-8 SuperNIC enables clustering of multiple Super AI Stations together and connection to external storage servers for high-speed data ingestion.

• Tried-and-true liquid cooling ingenuity:

  Inspired by Supermicro’s rich expertise in the development of data center cooling infrastructure, the closed loop liquid cooling architecture enables near-silent operation around the clock, while user serviceability reduces costly downtime and extends system longevity, minimizing interruption to always-on autonomous agents.

• Develop agentic AI locally with complete control:

  The Super AI Station is purpose-built for local AI development, empowering AI natives to exercise complete data sovereignty and total control over their security posture while minimizing external risks from storing and processing proprietary data on public clouds. Robust endpoint security features, including hardware root of trust, secure boot, and full air-gapped deployment capabilities ensure maximum data protection for deployments in even the most security-conscious environments. In addition, teams and neoclouds can also host multiple secure user instances simultaneously, with support for up to seven Multi-Instance GPU (MIG) instances on a single system.

• Future Windows Support:

  Supermicro deskside systems powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip will be coming to Windows later this year, empowering enterprise teams to build, run, and connect always-on, frontier AI agents to Windows applications and workflows.

All You Need to Develop and Deploy Powerful AI Agents

No matter where you are in your agentic AI development journey, the Super AI Station’s performance, flexibility, and robust security controls make it the perfect platform that can scale in line with your workloads. Backed by Supermicro’s proven system design expertise, rack-scale deployment capabilities, and comprehensive end-to-end portfolio the Super AI Station is ready to support AI natives and neoclouds throughout their AI factory journey, from their first deployment right up to a rack-scale NVL72 AI Factory.

Key Technical Specifications:

Join the NVIDIA Fast Start Program and Book Your Super AI Station Today

Kick off your AI development journey with NVIDIA’s Fast Start Program. Eligible neocloud operators and AI natives can enjoy up to 50% off their first 2 units of the Super AI Station. For more information and to order your Supermicro Super AI Station, reach out to your Supermicro representative or contact superstations@supermicro.com today!

For more information about Supermicro’s Super AI Station or to make a purchase, please visit the Super AI Station page. If you’re visiting COMPUTEX 2026, be sure to stop by Supermicro’s booth at Nangang Exhibition Center Hall 1, 4F, N0602 to experience exciting live demonstrations of the Super AI Station and speak with our product experts.