AI Infrastructure and Liquid-cooled Controls Design based on EcoStruxure Building Operation

AI factories and liquid cooling bring new challenges to data center design. Schneider Electric’s data center reference designs help shorten the planning process by providing validated, proven, and documented AI infrastructure designs to address such challenges. This design focuses on the deployment of AI infrastructure for AI factories, specifically for NVIDIA GB200 NVL72 and GB300 NVL72 based cluster such as NVIDIA DGX SuperPOD, in a single data hall. It will be extensible to NVIDIA Vera Rubin NVL72. The data hall is purpose-built and optimized using liquid-to-liquid CDUs and fluid coolers with adiabatic assist. Facility power and cooling design are optimized for capital cost, efficiency, and reliability.Controls Reference Design CRD2 includes information on these technical areas: EPMS/BMS interoperability with NVIDIA Mission Control (AI factory operations and orchestration software, including Cluster and Workload Management features) and other local or enterprise applications, Controls System for cooling infrastructure, coolant distribution units (CDUs), electrical infrastructure as Remote Power Panels (RPPs), end to end Electrical Distribution from high voltage to the low voltage power trains feeding IT and mechanical loads, including new guidance for measuring AI rack power profiles.