Mitsubishi Heavy Industries, Ltd. (MHI) has introduced strategic developments in infrastructure applied sciences to handle the increasing calls for of high-performance computing environments, notably AI Factory deployments. As these deployments speed up, infrastructure wants are evolving past conventional information facilities to industrial-scale environments.
Leveraging its experience in energy techniques, cooling applied sciences, and industrial engineering, MHI is advancing an built-in AI infrastructure portfolio. This portfolio combines superior cooling, modular techniques, and energy applied sciences to assist the transition to scalable AI-ready information middle infrastructure.
In a major commercialization milestone, MHI has shipped a 10MW-class centrifugal chiller check unit to the United States. Scheduled to reach on the Port of Brunswick, Georgia, the cargo is geared toward supporting next-generation high-density AI workloads.
As a part of its technique, MHI is creating the Modular Chiller Plant (MCP), a pre-engineered cooling system that integrates the centrifugal chiller with pumps, warmth exchangers, and controls right into a modular structure. The MCP helps NVIDIA DSX-aligned cooling architectures and scalable liquid-cooled AI deployments, simplifying deployment and enabling scalable enlargement.
MHI’s MCP structure is designed to boost Power Usage Effectiveness (PUE) and handle water shortage considerations with its closed-loop configuration, bettering Water Usage Effectiveness (WUE) in large-scale information facilities. The MCP is presently present process U.S. security and regulatory certification, together with UL certification, with the centrifugal chiller cargo contributing to certification readiness.
By decreasing onsite integration necessities and supporting future progress, the MCP helps hyperscale, colocation, and AI infrastructure operators handle cooling as computing densities enhance. Delivering AI factories at scale requires collaboration throughout compute, energy, and cooling sectors.
NVIDIA DSX, NVIDIA’s AI factory-scale platform, unifies design, simulation, operations, and ecosystem applied sciences. MHI participates within the NVIDIA Partner Network as a Power & Cooling Partner, working with NVIDIA and ecosystem companions to advance built-in energy and cooling applied sciences for NVIDIA DSX.
MHI’s options are underpinned by its experience in power, engineering, and industrial infrastructure techniques. The centrifugal chiller relies on confirmed expertise with a robust observe document in mission-critical functions, whereas MHI’s superior energy initiatives, together with 800VDC, are developed in collaboration with ecosystem companions.
Shin Gomi, Senior General Manager, Data Center & Energy Management at MHI, emphasised the significance of mixing superior cooling applied sciences, modular design, and industrial engineering experience with NVIDIA AI infrastructure to allow scalable, dependable, and energy-efficient AI infrastructure.
Vladimir Troy, Vice President of AI infrastructure at NVIDIA, highlighted the function of MHI’s work in large-scale cooling and 800 VDC energy infrastructure in serving to ecosystem companions construct extra scalable and energy-efficient AI factories.
Looking forward, MHI plans to increase its built-in AI infrastructure portfolio by expertise investments that assist the evolving necessities of AI Factory and high-density computing environments.