Introducing Supermicro Solutions featuring NVIDIA RTX PRO Blackwell Server Edition GPUs
Powerful, flexible multi-workload acceleration from AI factory to data center to edge
Supermicro’s modular and flexible accelerated solutions featuring NVIDIA RTX PRO BSE GPUs provide unprecedented performance and efficiency for AI, HPC, graphics, and video with industry-leading time-to-online. With a range of form factors and GPU configurations, Supermicro and NVIDIA solutions deliver up to 45x performance and 15x price/performance vs CPU-only compute, reducing infrastructure outlay and ongoing costs. With support for NVIDIA AI Enterprise software, full-stack solutions are right-sized for optimal multi-workload acceleration from AI factory to data center to edge.
Flexible and Modular with Industry-leading Time-to-Market
Supermicro’s broad selection of systems based on industry-standard PCIe GPU form factors allows easy customization for specific workloads and deployments.
Data Center Performance that can be Deployed Anywhere
Wide range of form factors supports wide-ranging deployment from AI factory to network edge, providing powerful and efficient acceleration throughout the AI compute pipeline.
Complete, Full-Stack Solutions with Supermicro & NVIDIA
Supermicro NVIDIA-Certified Systems™ that ensure compatibility with NVIDIA networking and software to create turnkey solutions that reduce the complexity.
NVIDIA RTX PRO™ Blackwell Server Edition GPUs are multi-workload accelerators designed to deliver industry-leading performance and efficiency for enterprise AI workloads, providing an easy upgrade path without redesigning racks, power, or cooling, harnessing Blackwell-class performance within existing environments. The NVIDIA RTX PRO 4500 Blackwell Server Edition delivers breakthrough performance for demanding data processing, AI, video, and visual computing workloads in a single‑slot, power‑efficient design, while the NVIDIA RTX PRO™ 6000 Blackwell Server Edition is a powerful data center GPU for AI and visual computing, accelerating demanding enterprise workloads including AI, scientific computing, graphics, and video applications.
| Feature |  NVIDIA RTX PRO 4500 BSE |  NVIDIA RTX PRO 6000 BSE |
|---|---|---|
| Target Workloads |
|
|
| Memory | 32GB GDDR7 @ 800 GB/s | 96GB GDDR7 @ 1597 GB/s |
| MIG Instances (per GPU) | Up to 2 @ 16 GB | Up to 4 @ 24 GB |
| Power | 165W (passive cooling) | 400-600W (passive cooling) |
| Form Factor | Single-width FHFL | Double-width FHFL |
| Single-Precision Performance | 51 TFLOPS | 120 TFLOPS |
| Peak RT Core Performance | 154 TFLOPS | 355 TFLOPS |
NVIDIA RTX PRO Blackwell Server Edition GPUs deliver significant performance and cost-performance increases over previous generation GPUs for a range of enterprise AI, HPC and visual computing workloads
Best Server for Enterprise AI
RTX Pro Server Up to 5x Better Price Performance
Performance/$ = Performance / TCO for a Single Node (Server + Power Costs) for L40S compared to RTX PRO 6000 Blackwell Server Edition
Performance based on NVIDIA testing
Up to 4x vs L4
3D Graphics, Rendering, AI-Content Creation
Performance based on NVIDIA testing
Supermicro AI Factory Systems
High-capacity GPU systems designed for AI factory and data center deployments, supporting physical AI, large-scale AI inference, enterprise HPC, and intensive graphics workloads. Architectures are thermally optimized for air-cooled environments, with select models also supporting high ambient temperatures of up to 35°C.
| Product(s) |  4U 8-GPU Available in direct-connect and dual-root configurations |  4U RTX PRO Server High-density NVIDIA MGX architecture |  5U 8-GPU Thermally-optimized design for efficient cooling and maximum storage |  3U 8-GPU Ultra-high-density configuration for on-prem AI inferencing |
|---|---|---|---|---|
| Key Applications | AI Factory, Large-scale AI Inference, FP32 HPC | AI Factory, Large-scale AI Inference | Large-scale AI Inference, FP32 HPC | AI Inference |
| GPU |
|
|
|
|
| CPU |
| Dual Intel Xeon 6900 Series | Dual Intel Xeon 6900 Series | Dual Intel Xeon 6900 Series |
| Storage |
|
| 2.5" NVMe | EDSFF E1.S |
Supermicro Enterprise Data Center Systems
Compact and efficient architectures designed for CPU-only compute replacement in existing data centers, providing balanced performance and acceleration for enterprise AI workloads including small model AI inference, AI development, media transcoding, multi-user graphics, and data analytics.
| Product(s) |  2U RTX PRO Server Compact dual processor design for AI factories |  2U Vera C2 High-density solution with energy-efficient NVIDIA Vera CPU |  2U Hyper Flagship performance rackmount platform in UP or DP architecture |  1U Hyper Flagship performance rackmount platform in DP architecture |
|---|---|---|---|---|
| Key Applications | Data Analytics, Visual Computing | Agentic AI, Virtualization, HPC | HPC, Virtualization | HPC, Virtualization |
| GPU |
|
|
| 1 NVIDIA RTX PRO 4500 Blackwell |
| CPU | Dual Intel® Xeon® 6700 Series | Dual NVIDIA Vera CPU |
| Dual Intel Xeon 6700 Series |
| Storage | 2.5" NVMe | EDSFF E1.S | 2.5" NVMe | 2.5" NVMe |
Supermicro Compact and Edge Systems
Compact and short-depth form factors designed to enable low latency at the network edge. These architectures maximize power and thermal efficiency, ideal for acceleration closer to the data source for time-critical workloads like edge AI inference, real-time analysis, and video streaming/media.
| Product(s) |  1U Compact Edge Server Short-depth chassis for edge data centers |  2U Compact Edge Server Short-depth chassis with front or rear I/O |  2U Hyper-E Flagship data center performance at the edge |  2U UP RTX PRO Server Single processor architecture optimized for edge AI inference with front or rear I/O |
|---|---|---|---|---|
| Key Applications | Edge AI, Enterprise Edge | Edge AI, Telco Edge | Edge AI, 5G Networking | Edge AI, Financial Services, Scientific Research |
| GPU | 1 NVIDIA RTX PRO 4500 Blackwell | 1 NVIDIA RTX PRO 4500 Blackwell | Up to 3 NVIDIA RTX PRO 4500 Blackwell |
|
| CPU | Single Intel Xeon 6700 Series | Single AMD EPYC 9005 series | Dual Intel Xeon 6700 Series | Single Intel Xeon 6700 Series |
| Storage | 2.5" NVMe | 2.5" NVMe/SATA | 2.5" NVMe | EDSFF E1.S |
Accelerating Enterprise AI with Supermicro and NVIDIA: RAG-ready Infrastructure
Harness the power of LLMs more securely and effectively for greater business impact with Supermicro and NVIDIA turnkey Retrieval-Augmented Generation (RAG) solutions. Unlock enterprise-ready AI by connecting LLMs directly to organization-specific data for precise, business-relevant results while maintaining strict data governance and privacy.
Deploy Production AI Faster with Supermicro NVIDIA-Certified Systems™
Take the guesswork out of creating complete AI solutions with Supermicro NVIDIA-Certified Systems validated for performance, manageability, security, and scalability.
- Guaranteed compatibility and support for NVIDIA AI Enterprise software to simplify the process of developing and deploying production AI.
- Scale-out compute nodes with PCIe-optimized Reference Configurations
- Supermicro NVIDIA-Certified and NVIDIA-Qualified systems available in 8-GPU, 4-GPU, and 2-GPU RTX PRO 6000 Blackwell Server Edition configurations
