Schneider Electric with NVIDIA to Develop Blueprints for Gigawatt-Scale AI Factories
Tuesday, March 17, 2026
Schneider Electric, an energy technology company, in collaboration with NVIDIA and industrial software company AVEVA, has announced key advancements in designing, simulating, building, operating and maintaining the next generation of AI datacentre infrastructure during NVIDIA GTC in San Jose. They include a new NVIDIA Vera Rubin reference design that validates power and cooling for the latest NVIDIA rack-scale architectures, integration of advanced digital twin capabilities within the NVIDIA Omniverse DSX Blueprint and ecosystem; and early testing of agentic AI for datacentre alarm management services using NVIDIA Nemotron open models. Contd
The company’s announcements further strengthen Schneider Electric and NVIDIA’s existing collaboration and establish a comprehensive foundation for developing AI Factories built for gigawatt-scale and efficiency.
New NVIDIA Vera Rubin Reference Design
The newly unveiled AI reference design is one of the first created for the NVIDIA Vera Rubin NVL72 racks. The validated reference design covers power and cooling and is integrated with Schneider Electric’s controls reference designs. Importantly, the design addresses important infrastructure requirements and considerations for NVIDIA’s latest rack-scale systems -
- Enables new power distribution with increased supply voltage of 480 VAC
- Allows higher TCS loop supply temperature of 45°C for enhanced efficiency
- Supports new IT room architecture with clusters of AI racks sharing centralised networking, storage, CPU and support racks. This allows every NVIDIA rack-scale system to remain physically close together while allowing separate, higher voltage for the GPU racks to enable larger clusters and optimise power delivery
- Maximises token performance by designing datacentrers to accommodate various operating points of GPU racks (both MaxP and MaxQ). Operating at MaxQ can achieve more tokens per watt to override any power constraints and optimise computing performance through redundancy. Overall, the reference design enables more tokens per watt when incorporating NVIDIA’s MaxQ operating point.
The reference design is validated with ETAP models for electrical system design and ITD CFD models for layout and air flow.
Additionally, AVEVA, a provider of industrial software owned by Schneider Electric, together with NVIDIA, has announced a new lifecycle digital twin architecture that maximises GPU efficiency and accelerates the deployment of AI factories at speed and scale. Schneider Electric is committed to creating SimReady assets and digital twins through NVIDIA Omniverse, supported by AVEVA’s advanced software. With this announcement, AVEVA’s engineering and operations software is now embedded throughout the NVIDIA Omniverse DSX Blueprint and ecosystem. It is projected to accelerate time-to-token through domain-specific simulations, digital visualisation and collaborative design tools that will drive significant engineering optimisation.
After a system architecture is assembled in the NVIDIA Omniverse environment, AVEVA executes multi-domain simulations to validate operational behavior under realistic conditions. This includes computational models for power distribution, thermal dynamics, airflow performance and controls. These simulations enable iterative design optimisation, rapid evaluation of multiple scenarios across a wide range of load and environmental conditions and final system verification prior to building the physical environment. The result is a fully validated, performance-optimised design that reduces engineering cycles and improves deployment accuracy.
“As AI workloads scale in both size and complexity, the margin for error in datacentre design becomes incredibly small,” said Manish Kumar, Executive Vice President, Secure Power & Datacentres at Schneider Electric. “Delivering AI at scale requires tightly integrated electrical, cooling and digital architectures that can support both unprecedented performance demands while maintaining peak energy efficiency. By combining advanced software, digital twins and validated reference designs, operators can simulate and optimise infrastructure before a single rack is deployed. This approach reduces risk, accelerates deployment and ensures the efficiency and resilience needed to power the next generation of AI factories.”
“Gigawatt-scale AI factories demand a fundamentally new class of energy-efficient and highly predictable infrastructure,” said Vladimir Troy, VP, AI infrastructure at NVIDIA. “Together, NVIDIA and Schneider Electric are providing the power, cooling and digital twin architectures needed to accelerate time-to-token for our customers worldwide.”
