Siemens: Helping Operators Build Future-Ready Data Centres

Siemens and nVent have unveiled a joint reference architecture tailored for hyperscale AI data centres, combining industrial power systems with liquid cooling technologies to support high-density Nvidia-based infrastructure.

The design is intended for facilities scaling to 100 megawatts (MW) and above, with a focus on modularity, deployment speed and energy efficiency.

The architecture centres around support for Nvidia’s DGX SuperPOD platforms, including the DGX GB200 system. These are purpose-built for AI workloads, requiring tightly integrated electrical and thermal environments to operate at full capacity.

Sara Zawoyski, President of nVent Systems Protection, says the collaboration reflects nVent’s long-term focus on thermal management.

“We have decades of expertise supporting customers’ next-generation computing infrastructure needs,” she adds. "This collaboration with Siemens underscores that commitment. The joint reference architecture will help data centre managers deploy our cutting-edge cooling infrastructure to support the AI buildout.”

Blueprint engineered for dense, liquid-cooled infrastructure

The design supports data centre campuses approaching 100MW, with AI-class compute workloads becoming more power-hungry and heat-intensive. Liquid cooling is a key part of the design, with nVent’s portfolio engineered to meet these challenges.

By aligning with Nvidia’s DGX SuperPOD reference design, Siemens and nVent aim to simplify integration for data centre operators, streamlining the roll-out of high-density environments. The model removes many of the fragmented engineering decisions that typically slow down projects, offering a pathway that combines power, cooling and automation in a single framework.

The reference architecture integrates Siemens’ power distribution systems, energy management platforms and scalable automation with nVent’s direct liquid cooling (DLC) and coolant distribution solutions. Siemens describes the model as “a blueprint for scale: modular, fault-tolerant, and energy-efficient". 

Ciaran Flanagan, Global Head of Data Center Solutions at Siemens, says: “This reference architecture accelerates time-to-compute and maximises tokens-per-watt, which is the measure of AI output per unit of energy. Together with nVent and our broader ecosystem of partners, we’re connecting the dots across the value chain to drive innovation, interoperability and sustainability, helping operators build future-ready data centres that unlock AI’s full potential.”

Tackling complexity in AI data centre growth

As AI infrastructure grows in scale and complexity, rack densities are rising fast. This puts pressure on both cooling systems and power infrastructure. Operators face challenges in areas like heat rejection, energy delivery and modular expansion, all of which are addressed in the Siemens–nVent architecture.

The blueprint is built to provide consistency from planning to operation, with standardised systems that help avoid delays caused by custom engineering and siloed deployment models. The integration also allows technology partners and hyperscalers to build around a unified foundation.

With liquid-cooled racks now playing a central role in AI deployments, maintaining electrical stability becomes critical. Siemens’ contribution ensures a reliable backbone, with systems that can monitor and regulate voltage across liquid-cooled environments and GPU-heavy clusters.

The focus is on enabling operators to build fast and scale reliably, while maintaining high energy performance. Tokens-per-watt, the efficiency metric mentioned by Siemens, becomes increasingly important in AI as it reflects the processing output relative to energy use—a key consideration in hyperscale environments.

Combining power resilience with thermal performance

Siemens offers a portfolio of industrial-grade tools, including medium and low voltage power distribution, IoT-connected hardware, energy analytics and software automation platforms. These systems support real-time transparency, improve operational control and reduce energy waste.

Digital services are also included to enable predictive maintenance and performance benchmarking across the entire data centre footprint. In high-density scenarios, where AI clusters require uninterrupted power, Siemens’ architecture maintains uptime through grid-interactive design and fault-tolerant distribution layers.

nVent brings expertise in high-density liquid cooling shaped through its partnerships with chip manufacturers, hyperscalers and server OEMs (original equipment manufacturers). Its systems support thermal efficiency at scale, without needing major changes when transitioning to future compute generations.

Its technologies include direct-to-chip cooling, coolant distribution units (CDUs) and management strategies designed for modular hall layouts. These are embedded into the blueprint, providing a ready-made model for fast deployment.

The architecture supports hyperscalers and operators aiming to meet AI demands without compromising on energy use or speed to market. Siemens and nVent position this model as a tool to help them "build the next wave of AI-ready data centres, where performance, sustainability and deployment speed are tightly linked to electrical design and advanced cooling".

By removing integration bottlenecks and aligning with Nvidia’s high-density standards, the architecture delivers a combined platform that supports operational goals at AI scale – allowing  infrastructure teams to plan for expansion without starting from scratch.