In the context of ISC 2025, MSI has unveiled its new range of high-performance enterprise server platforms designed to empower the most demanding artificial intelligence (AI) and accelerated computing workloads. From June 10 to 12, the company is showcasing at its booth #E12 solutions based on standardized modular architectures that promise to facilitate the rapid and scalable adoption of AI technologies in enterprise data centers of any size.
Danny Hsu, General Manager of MSI’s Enterprise Platform Solutions division, emphasized that “with modular and standards-based architectures, data centers can adopt artificial intelligence faster and more cost-effectively than ever. This marks the dawn of a new era where AI is not only powerful but also much more accessible for all types of businesses.”
Platforms for the Future of Enterprise AI
Based on the NVIDIA MGX modular architecture, MSI’s AI servers provide a flexible and powerful foundation for ever-evolving workloads. Notable among the new offerings is the CG480-S5063, a 4U server aimed at high-performance tasks such as training large language models (LLM), deep learning, and fine-tuning. This model supports dual Intel® Xeon® 6 processors and eight double-wide FHFL GPU slots, compatible with NVIDIA H200 NVL and NVIDIA RTX PRO 6000 Blackwell Server Edition, supporting GPUs of up to 600W. Additionally, it features 32 DDR5 DIMM slots and 20 PCIe 5.0 E1.S NVMe bays for exceptional memory bandwidth and data transfer speeds.
For inference tasks and lighter AI workloads in enterprise environments, the CG290-S3063 is a 2U platform also based on NVIDIA MGX, supporting up to 16 DDR5 DIMM slots, four GPUs of up to 600W, and flexible PCIe 5.0 storage. This model stands out for its scalable design, prepared to meet future industry needs.
The CX270-S5062, built under the DC-MHS (Datacenter Modular Hardware System) standard, is a 2U server with dual Intel Xeon 6 processors, aimed at critical applications such as virtualization and database management. It offers up to 32 DDR5 DIMM slots and 24 PCIe 5.0 U.2 NVMe bays for maximum performance in memory and storage.
For hyper-scale cloud environments, MSI presents the Open Compute CD281-S4051-X2, a 2OU, two-node server compliant with the OCP ORv3 standard. Each node integrates an AMD EPYC™ 9005 Series processor with a TDP of up to 500W, 12 DDR5 DIMM slots, and 12 PCIe 5.0 E3.S NVMe bays, along with air-cooling systems and efficient 48VDC power supply. It is designed for energy-efficient operations and large-scale deployments in next-generation data centers.
Scalability, Efficiency, and Future
MSI’s new platforms showcase a strategic shift in implementing artificial intelligence in enterprises: modularity, energy efficiency, and adaptability to the growing workloads associated with AI, deep learning, and cloud. These solutions are prepared to meet specific needs as well as support mass cloud deployments, anticipating the demands of a digital future.