MSI Expands AI Infrastructure Portfolio with Liquid-Cooled Racks and NVIDIA Partnerships
Event summary
- MSI showcased liquid-cooled AI infrastructure at COMPUTEX 2026, including ORv3 rack-scale architectures and NVIDIA MGX platforms.
- The company introduced a range of NVIDIA-based AI servers and desktop AI supercomputing platforms, including the XpertStation WS300.
- MSI's DC-MHS multi-node and enterprise server platforms support modular cloud and enterprise deployments.
- The liquid-cooled ORv3 rack architecture supports up to 100kW deployments with integrated Liquid-to-Liquid Coolant Distribution Unit (L2L CDU).
- MSI's NVIDIA MGX server portfolio includes 2U, 4U, and 6U GPU platforms designed for AI training, inference, and HPC workloads.
The big picture
MSI's expansion of its AI infrastructure portfolio comes as data centers face increasing thermal demands due to higher compute density. The company's focus on liquid-cooled solutions and partnerships with NVIDIA, AMD, and Intel positions it to capitalize on the growing demand for scalable AI training and inference platforms. The strategic shift towards modular and flexible deployment options reflects broader industry trends towards efficient and adaptable data center infrastructure.
What we're watching
- Thermal Efficiency
- The pace at which liquid-cooled infrastructure will become standard in high-density AI data centers.
- NVIDIA Partnership
- How MSI's collaboration with NVIDIA will impact the development of next-generation AI infrastructure.
- Market Adoption
- Whether enterprises will widely adopt MSI's modular cloud and enterprise server platforms.
Related topics