MSI Advances Next-Generation Data Center Infrastructure with AMD EPYC™ 9006 Series Server CPUs

Expanding AI, cloud, HPC, and enterprise capabilities through GPU, Core Compute, Open Compute, and enterprise server platforms Taipei, Taiwan – November 11, 2026 – As agentic AI drives new demands for compute density, memory bandwidth, and data movement, MSI is advancing next-generation data center infrastructure with server platforms powered by AMD EPYC™ 9006 Series SP7 Server CPUs. With SP7 systems entering the shipping phase in Q4 2026 and SP8 platforms planned for the first half of 2027, MSI’s portfolio spans liquid-cooled GPU servers, 19” DC-MHS and 21” ORv3 multi-node platforms, and enterprise servers, supporting AI, cloud, HPC, virtualization, and business-critical workloads. Built on the Zen 6 architecture, AMD EPYC 9006 Series Server CPUs deliver up to 256 cores and 512 threads per socket for high-density computing and AI agent execution. With up to 1.6TB/s of memory bandwidth and PCIe Gen6 connectivity, they support data-intensive workloads and high-speed I/O for GPUs, networking, and storage. “As agentic AI reshapes data center workloads, infrastructure needs to evolve beyond compute performance alone,” said Danny Hsu, General Manager of MSI’s Enterprise Platform Solutions. “MSI’s AMD EPYC 9006 server platforms give customers the flexibility to align system architecture with their performance, deployment, and operational requirements.” “Enterprise AI has moved from experimentation to production, and agentic AI will place even greater demands on compute, memory bandwidth, efficiency and system-level orchestration,” said Dan McNamara, senior vice president and general manager, Compute and Enterprise AI, AMD. “With 6th Gen AMD EPYC Server CPUs, AMD is helping customers scale enterprise AI with the performance and efficiency needed to support increasingly complex data pipelines, AI services and intelligent applications across the modern data center.” Liquid-Cooled GPU Servers for AI Host Infrastructure Designed for accelerator-rich AI infrastructure, the MSI CG681-S6093 6U liquid-cooled GPU server supports dual AMD EPYC 9006 Series Server CPUs and up to 8 liquid-cooled GPUs. With 32 DDR5 DIMM slots, PCIe Gen6 expansion, 400G networking, and 8 front-accessible E1.S NVMe drive bays, the platform provides the compute, memory, and I/O resources required for AI host-node deployments. High-Density Multi-Node Computing MSI’s 19” DC-MHS Core Compute platform delivers an air-cooled multi-node architecture for cloud, virtualization, and scale-out infrastructure. The SP7-based CD270-S4091-X2 features a 2U 2-node design with 16 DDR5 DIMM slots and 8 E1.S NVMe drive bays per node, supporting high-density compute within standard 19” rack environments. Rack-Scale Computing with Open Architecture MSI’s 21” ORv3 Open Compute platforms integrate DC-MHS multi-node architecture with 48Vdc busbar power distribution, supporting high-density compute through rack-level power delivery. Available in SP7 and SP8 configurations, the portfolio includes single-socket air-cooled designs and dual-socket air- and liquid-cooled options, up to 32 DDR5 DIMM slots per node, and E1.S or E3.S NVMe storage for diverse cloud, AI, and HPC deployments. CD281-S4091-X2: 2OU 2-node, 1P SP7, air-cooled CD282-S6091-X2: 2OU 2-node, 2P SP7, air-cooled CD182-S6091-X2: 1OU 2-node, 2P SP7, liquid-cooled CD281-S4081-X2: 2OU 2-node, 1P SP8, air-cooled Flexible Enterprise Computing MSI’s AMD EPYC 9006 Series SP8 enterprise servers offer balanced compute, memory, and I/O capabilities for virtualization, databases, storage, and business-critical workloads. Available in 1U and 2U single- and dual-socket configurations, the platforms support up to 32 DDR5 DIMM slots, PCIe Gen6 expansion, OCP 3.0 networking, and E3.S or U.2 NVMe storage options. CD281-S4091-X2: 2OU 2-node, 1P SP7, air-cooled CX172-S4081: 1U, 1P SP8 CX272-S4081: 2U, 1P SP8 CX172-S6082: 1U, 2P SP8 CX272-S6082: 2U, 2P SP8

SC26

MSI Showcases Open Data Center and AI Infrastructure at 2026 OCP Global Summit

San Jose, California – Oct 13, 2026 – MSI showcases its latest open data center and AI infrastructure at the 2026 OCP Global Summit (Booth #A49), led by new 21-inch ORv3 systems and DC-MHS modular building blocks, alongside NVIDIA MGX rack and GPU server platforms and the XpertStation WS300 desktop AI supercomputer on NVIDIA DGX Station architecture. The ORv3 systems address higher-density air- and liquid-cooled deployments, while the DC-MHS HPM portfolio extends modular platform options across AMD and Intel architectures. NVIDIA MGX rack and server platforms support scalable accelerated computing in the data center, complemented by the XpertStation WS300 for local AI development. “Infrastructure requirements are changing quickly as customers balance higher compute density, AI growth, and new cooling models,” said Danny Hsu, General Manager of MSI’s Enterprise Platform Solutions. “Our focus is to give them more flexibility through open standards and modular platforms that can evolve with their workloads.” 21-Inch ORv3 Systems Built on DC-MHS with AMD EPYC™ 9006 Series Server CPUs MSI is expanding its 21-inch ORv3 portfolio with new DC-MHS-based 2-node platforms powered by AMD EPYC 9006 Series Server processors, supporting up to 600W TDP per node. Available in 2OU air-cooled and 1OU direct liquid-cooled designs, the systems combine 48V busbar power, PCIe 6.0 connectivity, and memory speeds up to 8,000 MT/s with RDIMM and 12,800 MT/s with MRDIMM for next-generation data center deployments. Each node supports 2 PCIe 6.0 x16 expansion slots and front-access NVMe storage, providing high-speed I/O and flexible expansion for data-intensive workloads. Together with DC-SCM-based management, the platforms give operators a consistent ORv3 foundation while allowing cooling density to scale with deployment requirements. CD281-S4091-X2: A 2OU 2-node air-cooled platform with 16 DDR5 DIMM slots and 8 front E1.S PCIe 6.0 NVMe drives per node, providing higher local storage capacity for dense compute deployments while retaining standard air-cooled operation. The platform is suited for workloads including front-end web serving, cloud services, and content delivery infrastructure. CD182-S6091-X2: A 1OU 2-node direct liquid-cooled platform with 32 DDR5 DIMM slots and 4 front E1.S PCIe 6.0 NVMe drives per node, plus optional additional E1.S storage, enabling greater compute density and memory capacity within a smaller rack footprint. DC-MHS HPMs Across M-FLW, M-DNO and M-SDNO Architectures MSI’s DC-MHS HPM portfolio spans M-FLW, M-DNO Type 4, M-DNO Type 2, and M-SDNO form factors, with platforms supporting AMD EPYC 9006 Series Server processors and Intel® Xeon® 6/6+ processors. The broader HPM lineup gives system designers a common DC-MHS foundation across different board dimensions, processor platforms, and server configurations. AMD DC-MHS HPMs: D6082, D4091, and D4081 span M-FLW and M-SDNO designs for AMD EPYC 9006 Series Server processors. Across the portfolio, the platforms support PCIe 6.0, up to 32 DDR5 DIMM slots and MRDIMM speeds up to 12,800 MT/s, covering both full-width and density-optimized server designs. Intel DC-MHS HPMs: D5062, D3066, D3071, and D3061 span M-FLW, M-DNO Type 4, and M-DNO Type 2 form factors. The D3071 adds support for Intel Xeon 6900E+ processors with 12-channel memory and speeds up to 8,800 MT/s MRDIMM, while the broader portfolio supports PCIe 5.0 and OCP 3.0 networking across multiple server configurations. Scalable AI Infrastructure Across Rack, Server and Deskside Deployments Beyond its OCP portfolio, MSI brings a broader AI infrastructure lineup spanning NVIDIA MGX rack-scale systems, standalone NVIDIA MGX-based servers, and a desktop AI supercomputer on NVIDIA DGX Station platform. The portfolio addresses deployment needs from centralized accelerated computing infrastructure to local AI development, with liquid cooling, high-speed networking, and flexible GPU configurations supporting different scale and workload requirements. NVIDIA MGX Rack: A 48RU rack configuration with 4 MSI CG681-S6093 liquid-cooled NVIDIA MGX systems, NVIDIA Spectrum-4 SN5600 data switching, and SN2201 out-of-band networking, providing an integrated foundation for rack-scale accelerated computing deployments. CG681-S6093: A 6U liquid-cooled NVIDIA MGX platform supporting up to 8 NVIDIA RTX PRO 6000 Blackwell Server Edition Liquid Cooled GPUs, with dual AMD EPYC 9006 Series Server processors, eight 400G NVIDIA ConnectX-8 ports, PCIe 6.0 expansion, and up to 32 DDR5 DIMM slots for high-density AI and accelerated computing workloads. CG290-S3063: A 2U NVIDIA MGX platform supporting up to 4 GPUs, including NVIDIA H200 NVL, NVIDIA RTX PRO 6000, and NVIDIA RTX PRO 4500 Blackwell Server Edition, providing a compact option for GPU-accelerated AI deployments where rack density and configuration flexibility are priorities. XpertStation WS300: A desktop AI supercomputer powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, combining up to 748GB of coherent memory with dual 400GbE networking via NVIDIA ConnectX-8 SuperNIC. With support for both Linux and Windows environments, the system gives developers greater flexibility across familiar AI development workflows. Two XpertStation WS300 systems can be clustered to scale AI workloads beyond a single system, supporting local prototyping, fine-tuning, and inference of large-scale AI models and agents. .article-Content .card__img img { height: auto; }

MSI Brings NVIDIA MGX Rack-Scale AI Infrastructure to 2026 OCP Global Summit

San Jose, California – Oct 12, 2026 – MSI showcases its latest AI infrastructure at the 2026 OCP Global Summit (Booth #A49), featuring NVIDIA MGX rack-scale systems, liquid-cooled GPU servers, and the XpertStation WS300 desktop AI supercomputer based on NVIDIA DGX Station platform. The lineup spans rack-scale accelerated computing, high-density GPU deployments, and local AI development, with NVIDIA networking and liquid cooling supporting AI workloads across different deployment scales. “As AI infrastructure scales, customers need more than GPU performance alone. Cooling, networking, and system design all have to work together,” said Danny Hsu, General Manager of MSI’s Enterprise Platform Solutions. “Our focus is to provide practical deployment options that help customers scale AI infrastructure from individual systems to rack-level environments.” NVIDIA MGX Rack Infrastructure for Accelerated Computing MSI’s 48RU rack based on NVIDIA MGX architecture integrates four CG681-S6093 liquid-cooled NVIDIA MGX systems with NVIDIA Spectrum-4 SN5600 data switching and SN2201 out-of-band networking. The configuration brings compute, networking, and liquid cooling together at rack scale to support high-density accelerated computing deployments. Each CG681-S6093 is a 6U liquid-cooled NVIDIA MGX-based platform supporting up to 8 NVIDIA RTX PRO 6000 Blackwell Server Edition Liquid Cooled GPUs, dual AMD EPYC™ 9006 Series Server processors, up to eight NVIDIA ConnectX-8 SuperNICs with 400Gbps Ethernet connectivity per port, PCIe 6.0 expansion, and up to 32 DDR5 DIMM slots. The platform is designed for AI and accelerated computing environments where GPU density, network bandwidth, and thermal management need to scale together. NVIDIA MGX-based Servers for Flexible AI Deployments For deployments that require a more compact footprint, the CG290-S3063 provides a 2U NVIDIA MGX platform supporting up to 4 GPUs, including NVIDIA H200 NVL, NVIDIA RTX PRO 6000, and NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs. Its 2U design gives data center operators a flexible option for GPU-accelerated workloads where rack density and GPU configuration requirements vary. Together, the 6U CG681-S6093 and 2U CG290-S3063 extend MSI’s NVIDIA MGX portfolio across different GPU densities, cooling approaches, and deployment requirements. XpertStation WS300 Extends AI Computing to the Desk Based on NVIDIA DGX Station platform, the XpertStation WS300 is a desktop AI supercomputer powered by the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip. The system combines up to 748GB of coherent memory with dual 400GbE networking via NVIDIA ConnectX-8 SuperNIC for local AI prototyping, fine-tuning, and inference. With support for both Linux and Windows environments, the system gives developers greater flexibility across familiar AI development workflows. Two XpertStation WS300 systems can also be clustered to scale AI workloads beyond a single system, extending local AI development capacity while maintaining a deskside deployment model. .article-Content .card__img img { height: auto; }

OCP Global 2026

MSI XpertStation WS300 Based on NVIDIA DGX Station Now Available

Taipei, Taiwan – Aug, 27 – MSI today announced that the XpertStation WS300 based on NVIDIA DGX Station architecture is now shipping. Powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, the deskside AI supercomputer is now available through ASI, D&H, and Newegg, bringing data-center-class AI capabilities closer to enterprise AI development and deployment. "Enterprises don't just need powerful AI systems; they need a straightforward way to bring them into production," said Danny Hsu, General Manager of MSI's Enterprise Platform Solutions. "XpertStation WS300 is designed around that reality, a system that's ready to run demanding AI workloads the moment it arrives, not one that requires months of additional integration." "AI is moving from experimentation to always-on, agentic workloads that demand powerful computing close to where developers and enterprise data reside,” said Allen Bourgoyne, director of Enterprise Products at NVIDIA. “With XpertStation WS300 based on NVIDIA DGX Station architecture, MSI is bringing NVIDIA Grace Blackwell Ultra computing to the deskside, giving enterprises the performance and scale to develop, customize and run advanced AI models and agents locally." Enterprise AI Built for Local Development The XpertStation WS300 is powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip and combines up to 748GB of coherent memory with dual 400GbE networking via NVIDIA ConnectX-8 SuperNIC for clustering two XpertStation WS300 systems for AI workload scaling. Designed for large-scale AI development, fine-tuning, multimodal AI, and high-performance local inference, the platform keeps larger foundation models resident in coherent memory, reducing CPU-GPU data movement and improving inference efficiency. As AI workloads evolve from chatbot-style interactions to large language models (LLMs) and autonomous AI agents, enterprises are facing higher inference demand, larger model footprints, and continuous AI execution. These changes are accelerating the shift toward on-premises AI infrastructure, enabling organizations to reduce recurring cloud inference costs, retain control of proprietary models and data, and continuously customize AI models for enterprise use. Extending beyond AI model development, the XpertStation WS300 also supports NVIDIA NemoClaw with the NVIDIA OpenShell runtime, enabling policy-controlled autonomous AI agents to operate securely within enterprise environments and helping organizations move from AI-assisted workflows to agentic AI deployment.

MSI Extends XpertStation WS300 and EdgeXpert Platforms with Support for NVIDIA Nemotron 3.5 Lightning and NeMo Switchyard

Local AI systems that enable intelligent model routing for AI agents San Jose, California – Aug 11, 2026 – MSI today announced that its XpertStation WS300 and EdgeXpert platforms now support NVIDIA Nemotron 3.5 Lightning, a lightweight open model purpose-built for agentic AI and NeMo Switchyard, an open-source model routing library. Switchyard supports systems of models by routing each workflow step to the best-suited model, such as frontier-scale models for complex reasoning, lightweight models like Nemotron 3.5 Lightning for high-volume specialized tasks, improving accuracy, latency, and cost. With this support, XpertStation WS300, based on the NVIDIA DGX Station platform, serves as a deskside AI supercomputer for local AI development,, fine-tuning, and multi-model agent workflows, while EdgeXpert, powered by NVIDIA DGX Spark, applies the same routing intelligence in a compact system for always-on AI agents. "No single system fits every stage of an AI agent's lifecycle," said Sam Chern, MSI Vice President of Marketing. "By supporting NVIDIA's latest routing and model technologies across both XpertStation WS300 and EdgeXpert, we're giving customers a complete path, from model training and fine-tuning to continuous agent deployment, without switching vendors or ecosystems.” XpertStation WS300: Scaling Multi-Model Workflows Within Each Node MSI XpertStation WS300 is built on the NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip, with up to 748GB of coherent memory. This capacity allows Switchyard to run alongside multiple models locally, including both frontier-scale models and Nemotron 3.5 Lightning, without offloading any step to external infrastructure. Two WS300 units can be connected directly via an NVIDIA ConnectX-8 SuperNIC for larger workloads, while PCIe Gen5/Gen6 NVMe storage keeps data pipelines fed during sustained, multi-model workloads. EdgeXpert: Scaling Agentic AI Across Multiple Nodes MSI EdgeXpert based on the NVIDIA DGX Spark platform, is powered by the NVIDIA GB10 Grace Blackwell Superchip, delivering up to 1 petaFLOP of AI compute with 128GB of coherent unified memory in a compact form factor. That footprint is enough to keep Nemotron 3.5 Lightning and models up to 200 billion parameters, running continuously on-device, so specialized, high-volume agent tasks stay local rather than routing out to the cloud. NVIDIA ConnectX-7 networking allows up to four EdgeXpert units to cluster together, extending Switchyard's routing across a larger pool of local models as agentic workloads grow. NVIDIA NeMo Switchyard and Nemotron 3.5 Lightning NVIDIA NeMo Switchyard is an open-source model routing library for AI agents. It evaluates each workflow step and automatically selects the most suitable available model, with no training or manual configuration required, and improves its routing decisions the more agents run. NVIDIA Nemotron 3.5 Lightning is a lightweight open model distilled from NVIDIA's frontier Nemotron 3 Ultra. Built on a hybrid mixture-of-experts architecture with 30 billion total parameters and 3 billion active parameters, it delivers fast, efficient inference and supports context windows of up to 1 million tokens, making it well suited to run continuously alongside larger models as part of a complementary system of models.

MSI Launches 6th Gen AMD EPYC™ Server Portfolio

Introducing new GPU, multi-node, and enterprise platforms designed to accelerate AI, cloud, HPC, virtualization, and business-critical workloads San Francisco, California – Jul 23, 2026 – MSI is expanding its server portfolio with platforms advanced by 6th Gen AMD EPYC™ Server CPUs across SP7 and SP8 designs to modernize data centers for AI, cloud, HPC, virtualization, and business-critical workloads. SP7 platforms deliver maximum density and performance with up to 256 cores and 512 threads per processor for performance-driven deployments, while SP8 platforms offer up to 128 cores and 256 threads per processor for balanced mainstream deployments. MSI’s latest portfolio spans liquid-cooled GPU servers, 19” and 21” DC-MHS multi-node platforms, and enterprise servers. At AMD Advancing AI 2026, MSI will showcase the CD270-S4091-X2 multi-node server as part of its latest 6th Gen AMD EPYC server lineup. “Data center architecture is becoming more workload-specific, and customers need systems that can be selected with purpose,” said Danny Hsu, General Manager of MSI’s Enterprise Platform Solutions. “MSI’s latest server portfolio powered by 6th Gen AMD EPYC Server CPUs gives organizations the platform range to build around performance, density, efficiency, and serviceability as their infrastructure evolves.” “AMD enterprise AI technologies have become the trusted foundation for the modern data center, and 6th Gen AMD EPYC Server CPUs extend that leadership with a new level of performance, efficiency and scalability,” said Dan McNamara, senior vice president and general manager, Compute and Enterprise AI, AMD. “As enterprises scale cloud, HPC, virtualization and AI workloads, EPYC CPUs provide a trusted foundation for AI performance, infrastructure consolidation and innovation, while continuing to deliver leadership general-purpose compute for the business-critical applications and workloads that keep organizations running every day.” Liquid-Cooled GPU Server With the CG681-S6093 Liquid-Cooled 6U GPU Server, MSI advances SP7-based AI infrastructure for dense GPU deployments with strong AI host-node capability and liquid-cooling efficiency. The 6U platform supports dual 6th Gen AMD EPYC Server CPUs, 400G networking, and E1.S NVMe storage to scale accelerated AI workloads. Core Compute Platforms MSI’s Core Compute platforms provide 19-inch DC-MHS multi-node architectures optimized for cloud, virtualization, containerized workloads, and scale-out enterprise applications. Available in both SP7 and SP8 configurations, these air-cooled platforms allow customers to balance compute density, performance, and power efficiency across modern data centers. CD270-S4091-X2: 2U 2-Node (1P SP7, air-cooled) for performance-driven cloud and enterprise deployments CD270-S4081-X4: 2U 4-Node (1P SP8, air-cooled) for cost-efficient scale-out cloud infrastructure Open Compute Platforms MSI’s Open Compute platforms bring 21-inch multi-node designs to ORv3 rack-scale infrastructure for cloud, HPC, AI cluster infrastructure, and scale-out enterprise workloads. Available in SP7 and SP8 designs, the platforms offer flexible socket and cooling configurations, including single-socket, dual-socket, air-cooled, and liquid-cooled options. These systems help customers build dense, serviceable compute infrastructure for large-scale deployments. CD281-S4091-X2: 2OU 2-Node (1P SP7, air-cooled) CD181-S4091-X2: 1OU 2-Node (1P SP7, liquid-cooled) CD282-S6091-X2: 2OU 2-Node (2P SP7, air-cooled) CD182-S6091-X2: 1OU 2-Node (2P SP7, liquid-cooled) CD281-S4081-X2: 2OU 2-Node (1P SP8, air-cooled) Enterprise Servers MSI’s Enterprise Servers are designed for mainstream enterprise IT, virtualization, storage, database, and business-critical workloads, with 1U and 2U SP8 platforms available in both single-socket and dual-socket configurations. Featuring versatile E3.S or U.2 storage configurations, OCP 3.0 networking, and expansion flexibility, these systems help customers scale reliable enterprise infrastructure across diverse workloads. CX172-S4081 (1U, 1P SP8) CX272-S4081 (2U, 1P SP8) CX172-S6082 (1U, 2P SP8) CX272-S6082 (2U, 2P SP8) AMD, the AMD arrow logo, EPYC, and combinations thereof are trademarks of Advanced Micro Devices, Inc.