Schneider Electric’s RapidScale Modular Data Center Reference Design 115 is a validated, pre-engineered blueprint for high-density AI infrastructure. The company says the design streamlines onsite logistics and planning while speeding deployment of AI factories worldwide.

The design pairs the NVIDIA Vera Rubin platform with liquid cooling technology intended to maximize processing power while keeping energy use under control. Advanced fluid cooling with smart temperature controls is meant to deliver peak compute performance without overheating or wasted power.

By building modularity into the broader infrastructure architecture, the approach steers operators toward repeatable, risk-tested AI infrastructure blueprints rather than relying solely on custom-engineered data centers. The blueprints combine power, liquid cooling, IT, software and lifecycle services into a scalable, industrialized delivery model.

Market shift drives demand for prefabrication

The announcement arrives as AI processing needs to reshape data center requirements. Schneider Electric estimates global data center capacity will grow by 170 GW by 2030, reaching 300 GW in total, or 2.3 times the 2025 installed base. By the end of the decade, AI workloads are expected to account for approximately 65% of total demand.

In that environment, the traditional benefits of prefabricated modular infrastructure—speed, predictability, repeatability and quality—have become essential requirements for AI-scale deployment, according to the company. Developers relying on traditional stick-built construction face growing difficulty keeping pace, while high-density computing intensifies power and cooling demands, complicating system integration and adds pressure on grid capacity, energy efficiency and regulatory compliance. Skilled labor shortages compound the problem by creating workforce bottlenecks and stretching installation timelines.

Compass datacenters study quantifies the gains

Separately, Schneider Electric and hyperscale developer Compass Datacenters co-authored a research paper, “Prefabricated White Space Module Time Study: Lessons from a High-Density Liquid-Cooled Deployment” (White Paper 238). It examines the efficiency gains of moving from traditional white space builds to prefabricated modular IT pods.

According to the study, performed at a Compass hyperscale data center campus, prefabricated modular pods shortened onsite installation time from 75.5 hours to 15 hours per pod, an 80% reduction. Shifting critical assembly, integration and verification into a controlled factory environment lets operators reduce onsite labor and complexity while bringing AI compute capacity online faster.

“The white space brings together the greatest concentration of systems and trades in the smallest footprint, with each installation dependent on the one before it,” said Adil Attlassy, Chief Technology Officer of Compass Datacenters. “Working jointly with Schneider Electric, we engineered a prefabricated module that integrates these systems into a single assembly, eliminating the need to coordinate multiple scopes of work onsite. By reducing onsite installation time by 80%, we accelerate deployment and help our customers bring their workloads online sooner.”

An industrialized architecture for the AI lifecycle

RapidScale is Schneider Electric’s end-to-end approach to designing, delivering and supporting prefabricated modular AI infrastructure at scale. It draws on engineering across power, cooling, IT space, software and lifecycle services to help customers move away from complex custom engineering and toward repeatable blueprints. The new reference design is also a foundational first step toward supporting the NVIDIA DSX POD reference design, which enables factory integration and validation of Vera Rubin AI infrastructure before shipment to accelerate time to first token.

Supply chain management, the company noted, is now a critical factor in reducing scheduling uncertainty and project risk. Schneider Electric points to a globally integrated supply chain spanning individual components to complete modular solutions. Its manufacturing capacity has grown fourfold since 2025 and now covers eight primary sites totaling 1.35 million square feet, plus three additional facilities and 12 partner sites across 10 countries.

RapidScale incorporates approximately 90% in-house manufactured equipment, which the company says ensures predictable performance, supply chain certainty and full lifecycle support from day-one deployment through long-term optimization.

“As AI infrastructure scales at unprecedented speed, customers need solutions that are standardized, reliable, and deployable with certainty,” said Manish Kumar, executive vice president, secure power & data centers at Schneider Electric. “Through our portfolio of reference designs and prefabricated solutions, combined with global supply chain strength and lifecycle expertise, we help AI factory operators accelerate time to value while building the resilient infrastructure required for next-generation compute.”

Standards and availability

The reference design has been validated to ANSI standards for U.S. deployments, and Schneider Electric plans to extend the framework to support IEC standards for global implementations in the future.