Mitsubishi Electric has introduced a new "Chip-to-Grid" blueprint aimed at supporting the power and cooling requirements of AI factories built around NVIDIA accelerated computing infrastructure.
The company said the approach brings together technologies spanning power generation and grid infrastructure to data centre electrical systems and chip-level cooling. It is designed to address the increasing power density and thermal management requirements associated with large-scale AI workloads.
Under the blueprint, Mitsubishi Electric plans to integrate its portfolio of power and cooling technologies with NVIDIA's AI infrastructure. The approach covers multiple layers of an AI factory, including grid connections, electrical distribution, uninterruptible power supplies, power electronics and thermal management.
AI factories are data centre environments designed to support the development and operation of artificial intelligence models and applications. The increasing deployment of GPU-based computing systems has placed greater demands on data centre power capacity and cooling infrastructure, prompting infrastructure providers to explore more integrated approaches to facility design.
Mitsubishi Electric's blueprint seeks to address these requirements by treating power delivery and cooling as connected parts of the infrastructure rather than independent systems. This includes managing electricity from the grid through the data centre and ultimately to computing hardware, while coordinating cooling systems required for high-density AI servers.
The company said the architecture is intended to help data centre operators improve energy efficiency, reliability and scalability as AI computing requirements increase. Mitsubishi Electric is also looking at technologies that can monitor and optimise energy consumption across AI factory infrastructure.
The announcement expands Mitsubishi Electric's involvement in the data centre market, where the company provides equipment and technologies across power management, electrical infrastructure and cooling. The company has also been increasing its focus on infrastructure designed for high-density computing environments.
NVIDIA has been promoting the concept of AI factories as organisations invest in infrastructure capable of training and running increasingly complex AI models. These facilities combine accelerated computing, networking, software and supporting infrastructure to produce AI services at scale.
Mitsubishi Electric's Chip-to-Grid approach reflects a wider shift in data centre development as infrastructure companies respond to the energy requirements created by accelerated computing. Power availability, electrical distribution and cooling capacity are increasingly becoming considerations in determining where and how new AI infrastructure can be deployed.
The company is positioning the blueprint as an integrated framework for operators planning new AI factories or expanding existing facilities as demand for AI computing capacity continues to grow.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.