Schneider Electric, NVIDIA Co-Engineer GB300 NVL72 Power and Cooling Frameworks

Author photo: Craig Resnick
ByCraig Resnick
Category:
Acquisition or Partnership

New controls and power management reference designs co-engineered with NVIDIA accelerate deployment of AI-ready data center infrastructure.

Schneider Electric announced new reference designs developed with NVIDIA to help accelerate time to deployment for AI-ready infrastructure solutions. The new controls reference design provides an integrated framework for power management and liquid cooling controls, enabling more seamless IT/OT interoperability with NVIDIA Mission Control and enterprise applications. A second reference design supports the deployment of NVIDIA GB300 NVL72-based clusters, co-engineered with NVIDIA to help operators adopt next-generation Blackwell GPUs. Together, these reference designs provide data center operators with a faster path to deploying AI infrastructure for AI factories at scale.

Integrated Controls Reference Design

The controls reference design delivers a standardized framework for integrated power and liquid cooling management, including Motivair by Schneider Electric liquid cooling technologies. It features more seamless interoperability with NVIDIA Mission Control for cluster and workload management and can also be applied to Schneider Electric’s existing data center reference designs for NVIDIA Grace Blackwell systems.

Key Capabilities

  • A standardized interface publishing power management and liquid cooling controls data for local applications and upstream systems, such as AI infrastructure management software, digital twins, and enterprise tools

  • Redundancy management across cooling and power distribution infrastructure, including coolant distribution units (CDUs) and remote power panels (RPPs), to help ensure resiliency

  • New guidance for measuring AI rack power profiles, with a focus on peak power and power quality monitoring

NVIDIA GB300 NVL72 Reference Design

The NVIDIA GB300 NVL72 reference design supports the deployment of AI infrastructure with rack densities up to 142 kW, such as NVIDIA DGX SuperPOD with DGX GB300 systems. The design includes detailed frameworks for facility power, facility cooling, IT space, and lifecycle software, available in configurations for both ANSI and IEC standards.

Schneider Electric’s reference design also integrates ETAP and EcoStruxure IT Design CFD models, enabling users to simulate power and cooling scenarios and further optimize their data hall layout. This builds upon previous NVIDIA GB200 NVL72 blueprints, continuing Schneider Electric’s collaboration with NVIDIA to deliver engineered, validated models for next-generation accelerated computing environments.

Enabling Faster AI Factory Deployment

By providing validated, documented data center designs, Schneider Electric helps operators to design and implement critical infrastructure before the latest GPU systems arrive, helping to reduce deployment time and better optimize for efficiency and reliability. These reference designs form a foundation for AI factory architectures, helping operators to meet rising AI demand with more robust, scalable, and energy-efficient infrastructure.

Follow ARC Advisory Group for the latest trends in sustainability, automation, and next generation technology.

Engage with ARC Advisory Group

Representative End User Clients
Representative Automation Clients
Representative Software Clients