Schneider Electric today pulled back the curtain on a new set of AI data center reference designs created in partnership with NVIDIA. The goal is to make it faster and easier for operators to roll out what both companies are calling the infrastructure for “AI factories.” These blueprints focus on two key areas: integrated power and cooling controls, and support for NVIDIA’s latest high-density GPU racks built around the Blackwell architecture.
The first design is billed as the industry’s only critical framework that ties together liquid cooling and power management under one controls system. It’s compatible with NVIDIA Mission Control, the company’s orchestration platform for AI workloads, and is intended to bridge the divide between operational technology and IT. Schneider says this “plug-and-play” approach can help operators manage everything from coolant distribution units to rack-level power, all while ensuring redundancy, uptime, and real-time performance monitoring.
The second design addresses the NVIDIA GB300 NVL72 platform, which packs up to 142kW per rack and can scale to clusters of more than 1,100 GPUs. Schneider’s reference model includes technical guidance for power, cooling, space planning, and lifecycle management under both ANSI and IEC standards. By integrating digital twins through tools like ETAP and EcoStruxure, operators can simulate different scenarios before deploying.
Both designs highlight a shift toward standardization and speed in AI infrastructure deployment. With GPU demand climbing and data center density reaching new levels, Schneider Electric and NVIDIA are positioning these frameworks as roadmaps that let operators build quickly without compromising reliability or efficiency. Pricing wasn’t disclosed, but the company emphasized that these designs are available now for operators looking to stay ahead of the AI curve.
