Mitsubishi Electric's U.S. subsidiary, Mitsubishi Electric Power Products, has released Chip-to-Grid DSX Reference Designs to support adoption of NVIDIA's Vera Rubin NVL72, NVIDIA's next-generation rack-scale AI supercomputing system that integrates 72 Rubin GPUs and 36 Vera CPUs, along with future NVIDIA AI infrastructure deployed in U.S. data centers. Going forward, Mitsubishi Electric plans to use these reference designs as the basis for developing solutions aimed at next-generation AI data centers.
The Chip-to-Grid DSX Reference Designs cover and coordinate the full chain of systems involved in running a data center, from the connection to the electrical grid down to power delivery at the chip level, as well as cooling that spans from individual server racks up to the entire facility. This comprehensive approach combines Mitsubishi Electric's established power transmission and distribution equipment with next-generation power supply technology and cooling systems, and is designed to offer reliable scalability to support future data centers consuming gigawatt-scale power.
Specifically, data center sites built around this approach would be equipped with on-site power generation and battery energy storage systems (BESS). During normal operation, this setup is meant to minimize the site's reliance on the external power grid, and in the event of a grid failure, the facility would switch to running autonomously on its own microgrid, helping stabilize both power and operations. The designs also support both conventional 415/480VAC power distribution and the newer 800VDC standard within the data center. Combined with a dual-loop cooling system, an architecture that pairs direct liquid cooling of the chip itself with air cooling for other components, the overall solution is intended to maximize power efficiency for high-density AI workloads.
Mitsubishi Electric frames this solution for optimizing data center power operations as one of its key strategic priorities. In recent years, AI use has expanded beyond generative AI, which has already been widely adopted, into agentic AI and physical AI. As a result, data centers are expected to grow rapidly as critical infrastructure, but operators need to balance that growth with high power efficiency, reliability, and scalability. Data center operators face significant challenges around securing enough power, managing heat, and maintaining stable operations as AI demand keeps increasing.
Going forward, Mitsubishi Electric plans to use these new reference designs to help address these data center challenges and expand its business globally by offering high-value solutions that integrate with its broader range of data center systems. Through this work, the company aims to support the AI infrastructure essential to future societies, while contributing to its own sustainable growth.
