Supermicro Now Accepting Orders on Portfolio of More Than 20 Systems Optimized for the New NVIDIA RTX PRO 6000 Blackwell Server Edition PCIe GPUs, Accelerating the Deployment of Enterprise AI Factories

• More than 20 systems are available with NVIDIA RTX PRO 6000 Blackwell GPUs including NVIDIA-Certified systems
• Supermicro collaborating on development of new NVIDIA Enterprise AI factory validated designs based on RTX PRO Servers and NVIDIA HGX™ B200 systems
• Supermicro's new 4-GPU system based on the NVIDIA MGX™ reference design, bringing NVIDIA RTX PRO Server closer to the Edge for more powerful AI inference

Taipei, Taiwan. – May 19, 2025 – Supermicro, Inc. (NASDAQ: SMCI) a Total IT Solution Provider for AI/ML, HPC, Cloud, Storage, and 5G/Edge today announced that it is now taking orders for Enterprise AI systems with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. Supermicro's broad portfolio of optimized servers enables AI and visual computing to be deployed in virtually any industry or environment, with the NVIDIA RTX PRO Blackwell Server Edition GPU significantly enhancing performance for Enterprise AI workloads including AI inference & fine-tuning, AI development, generative AI, AI-driven graphics & rendering, video content & streaming, and game development. Supermicro NVIDIA-Certified systems with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs will serve as the building blocks for Enterprise AI Factories, integrating with Spectrum-X networking, NVIDIA-Certified Storage, and NVIDIA AI Enterprise software to create full-stack solutions, accelerating the deployment of on-premises AI.

"Supermicro continues to lead the development of enterprise AI infrastructure, empowering the deployment of AI across industries at ever-greater scale," said Charles Liang, president and CEO of Supermicro. "Supermicro's Data Center Building Block Solutions® is the ideal platform for collaboration with NVIDIA on Enterprise AI Factory validated designs based on the Blackwell architecture. Together, we will help enterprises ramp up AI adoption by building their own Enterprise AI Factories, accelerating AI inference, AI development, simulation, and graphics workloads for faster time-to-revenue."

For more information, please visit www.supermicro.com/pcie-gpu

Supported Supermicro architectures include 5U and 4U, NVIDIA MGX, and edge GPUoptimized systems, SuperBlade® and GrandTwin® multi-nodes, 1U and 2U rackmounts, multi-processor systems, and workstations. In total, Supermicro now offers more than 100 servers supporting NVIDIA PCIe GPUs including RTX PRO 6000 Blackwell Server Edition, H200 NVL, H100 NVL, L40S, and L4, meaning systems can be easily adapted based on specific workload and deployment environment requirements. These include NVIDIA Certified systems which guarantee compatibility and support for NVIDIA Spectrum™ networking and NVIDIA BlueField3 SuperNICs as well as NVIDIA AI Enterprise software, simplifying the process of developing and deploying production AI.

"AI factories are helping enterprises make smarter, data-driven decisions and swiftly adapt to market changes," said Chris Marriott, Vice President of Enterprise Platforms at NVIDIA. "Supermicro's diverse portfolio of NVIDIA Blackwell systems offer enterprises adaptable infrastructure to efficiently support AI workloads that drive innovation across virtually every business function."

As demand for AI inference continues to shift closer to the source of the request, Supermicro's all-new single-socket, GPU-optimized architecture based on the NVIDIA MGX™ reference design brings the NVIDIA RTX PRO 6000 Blackwell Server Edition's powerful AI inference capabilities closer to the edge. Supermicro's first MGX-based system leveraging a single-socket architecture, the new SYS-212GB-NR combines up to 4 latest-generation GPUs with balanced compute performance, resulting in a powerful yet cost-effective edge AI solution. The system has been designed to meet the needs of enterprises looking to deploy dense GPU resources in decentralized locations where power, thermal, and cost challenges may limit the effectiveness of traditional GPU system architectures. By deploying NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, enterprises can reduce the number of servers required to support advanced AI edge inference, business intelligence & analytics, industrial automation, and retail applications, further reducing the cost of deploying AI at the edge.

Supermicro offers the industry's broadest range of systems for Enterprise AI, with the following system families now shipping with the NVIDIA RTX™ PRO 6000 Blackwell Server Edition GPUs:

5U PCIe GPU – Highly flexible, thermally optimized architectures designed to support up to 10 GPUs in a single chassis with air cooling (up to 8 RTX PRO 6000 Blackwell). Systems feature dual-socket CPUs and PCIe 5.0 expansion to facilitate high speed networking for scale-out clusters. Key workloads include AI inference and fine-tuning, 3D rendering, simulation, and cloud gaming.

MGX - GPU-optimized systems based on NVIDIA's modular reference design, with single or dual-socket configurations supporting up to 4 GPUs in 2U or 8 GPUs in 4U for a range of data center and edge AI, HPC, and graphics workloads.

3U Edge-optimized PCIe GPU – Compact form factor designed for edge data center deployments and supporting up to 8 double-width or 19 single-width GPUs per system. Key workloads include EDA, scientific modeling, and edge AI inference.

SuperBlade® - Density-optimized and energy-efficient multi-node architecture designed and optimized for AI Inference, HPC, EDA, Visual Computing, Edge AI, with up to 40 GPUs per 8U enclosure and up to 120 GPUs per rack.

Rackmount Workstation – Workstation performance and flexibility in a rackmount form factor, offering increased density and security for organizations looking to utilize centralized resources.