In response to the growing demand for computing power for artificial intelligence, Schneider Electric is launching two new data centre reference designs in collaboration with NVIDIA. Their aim is to reduce deployment times and facilitate the construction of infrastructures capable of supporting the latest generation of GPUs, including the Blackwell family of chips.
The partnership combines NVIDIA’s dominance in AI computing with Schneider Electric ‘s expertise in power and cooling management. In an era where the availability of computing power is becoming a key bottleneck in the development of artificial intelligence, standardisation and acceleration of physical infrastructure is becoming critical.
The first project on show is a control system scheme that is unique on the market and, for the first time, comprehensively integrates the world of operational technology (OT) – that is, power and liquid cooling – with the world of IT. The solution provides full interoperability with NVIDIA Mission Control software for managing clusters and workloads in so-called AI factories.
With a plug-and-play architecture based on the MQTT protocol, data centre operators gain standardised access to data from the entire infrastructure. This enables not only real-time monitoring, but also the creation of digital twins of the entire facility. In practice, this means that the performance, energy consumption and reliability of power and cooling systems can be precisely optimised, which is crucial when handling energy-intensive next-generation processors.
The second project focuses on the physical architecture for ultra-high-density AI clusters, specifically for server racks with NVIDIA GB300 NVL72 systems, where power consumption can reach 142 kW per single rack. It is a comprehensive plan that covers four key areas: power, facility cooling, IT space arrangement and infrastructure lifecycle management software.
Available in variants for ANSI and IEC standards, this design provides the foundation for building next-generation data centres. By providing validated and tested schematics, Schneider Electric allows operators worldwide to design and deploy infrastructure even before the latest AI chips are physically available, minimising risk and optimising costs.
This initiative is a logical step in the evolution of the market. As AI systems increase in complexity and energy requirements, reliance on off-the-shelf, proven designs is becoming a necessity rather than a choice. The new proposals from Schneider Electric and NVIDIA are an attempt to create a universal standard for the foundations on which the future of artificial intelligence is to be built.
