Schneider Electric has unveiled solutions that can build out power infrastructure for artificial intelligence data centers in as little as six months, using designs co-developed with Nvidia. With power demand at data centers surging as high-performance graphics processing units are deployed, the core approach is to manufacture power supply equipment in advance to shorten construction timelines and support stable electricity delivery.
Schneider Electric said on the 23rd that it had unveiled its next-generation Prefabricated Power Modules and Power Skids. Both products are built on a reference design for AI data centers co-developed with Nvidia.
As the markets for generative AI and high-performance computing expand, a data center's power supply capacity has emerged as a key competitive factor. The denser the GPUs used for AI computation, the more electricity they consume and the more heat they generate. Data center operators are therefore focusing on securing reliable power and cooling systems while cutting the time needed to build out facilities.
The reference design applied to these products targets Nvidia's GB300 NVL72 system. It supports power density of up to 142 kilowatts per rack and integrates power management with liquid cooling controls. The design allows power supply and cooling equipment to run efficiently even in environments packed with high-density AI servers.
Based on that design, Schneider Electric has standardized power equipment and shifted to building it in factories ahead of time. Rather than installing individual equipment at the data center site, operators can use prefabricated modules to shorten buildout time. The company plans to use the approach to serve facilities that require large power capacity, such as clusters for training AI models and large-scale GPU server deployments.
The power modules incorporate Galaxy VXL, Schneider Electric's uninterruptible power supply technology. The products come in three capacities: 2 megawatts, 2.25 megawatts and 2.5 megawatts. The 2.5-megawatt version offers the largest capacity among the single-container prefabricated power solutions the company has released to date.
Power capacity can also be expanded by adding modules. Data center operators can build to initial demand and then scale up power capacity in stages as they add GPU servers. Schneider Electric plans to supply the products not only to large hyperscale data centers but also to neoclouds, which provide AI-dedicated cloud services, and to colocation data centers that host servers for multiple companies.
An official at Schneider Electric said securing the necessary power infrastructure quickly and reliably is critical to scaling AI data centers. The official added that by applying the design co-developed with Nvidia, the company aims to help customers handle high-density AI environments and expand their equipment in line with shifting power demand.
The new products will be designed, manufactured and tested at the company's production facility in Sant Boi de Llobregat, near Barcelona, Spain, before shipping. Schneider Electric recently expanded the site to meet rising demand for modular data center infrastructure. The products will launch first in Europe, with supply widening to North America during 2026 and rolling out to other regions, including Asia-Pacific, thereafter.