MSI Expands Nvidia RTX PRO Server Lineup with 4U CG480-S5063 and 2U CG290-S3063

MSI (Micro-Star INT’L Co., Ltd.) announced updated NVIDIA GPU systems that join the Nvidia RTX PRO server portfolio, the 4U CG480-S5063 and the 2U CG290-S3063 AI server, built on the Nvidia MGX modular architecture.

Both platforms support the Nvidia RTX PRO 6000 Blackwell Server Edition GPU, giving enterprises a practical way to increase AI and simulation performance without the need for a complete data center redesign.

“MSI is among the first to deploy Nvidia RTX PRO Servers,” said Danny Hsu, GM, enterprise platform solutions, MSI. “With the RTX PRO 6000 Blackwell Server Edition GPU integrated into our MGX platforms, we provide enterprise systems built for real AI workloads and support the transition toward AI-ready data centers.”

Next-Gen GPU Platforms for AI Era
The new MSI servers accelerated by Nvidia RTX PRO 6000 Blackwell Server Edition GPUs, support PCIe Gen 5.0 and DDR5 memory, paired with high-speed NVMe storage to deliver the compute density, bandwidth, and throughput required for modern AI workloads.

• The 4U Nvidia MGX server CG480-S5063 is targeted at compute-intensive AI and HPC deployments. It supports a 2-8-5-200 CPU:GPU:NIC topology, dual Intel Xeon 6 processors, up to 8TB of DDR5 memory across 32 DIMM slots, and 20 PCIe Gen 5.0 E1.S NVMe bays. With capacity for up to 8x600W FHFL dual-width GPUs, including RTX PRO 6000 Blackwell Server Edition GPUs, this high-density platform is optimized for large-scale training, inference, and simulation workloads.

2U NVIDIA MGX server CG290-S3063

• The 2U Nvidia MGX server CG290-S3063 offers a more compact form factor tailored for inference, small model training, and fine-tuning. It features a 1-4-3 GPU topology with a single Xeon 6 processor, supports up to 16 DDR5 DIMM slots, and provides 4 rear PCIe 5.0 U.2 NVMe bays plus 2xM.2 slots for fast data access. With 4 FHFL dual-width GPU slots in a streamlined 2U design, the CG290-S3063 balances compute performance with efficient use of power and space, making it ideal for enterprises scaling AI workloads with practical density.

Universal AI Acceleration with Nvidia RTX PRO Servers
Accelerated by Nvidia RTX PRO 6000 Blackwell Server Edition GPUs, the company’s servers deliver universal acceleration across enterprise workloads:

• Agentic and reasoning AI: Enabling advanced autonomous systems with improved cost-efficiency compared with Nvidia H100 GPUs.
• Physical AI and Simulation: Driving robotics, digital twins, and synthetic data generation with faster, higher performance than Nvidia L40S-based systems.
• Flexible Enterprise Integration: Supporting Windows, Linux, and leading hypervisors on air-cooled PCIe-based x86 platforms, with enterprise-grade security, manageability, and serviceability.

Resources:
Watch the MSI 4U and 2U Nvidia MGX AI platform, built on NVIDIA accelerated computing to deliver the performance for tomorrow’s AI workloads.