TECH TALK design and thoughtful operations across several key areas. The good news is operators have several key levers to shape AI infrastructure that is both high-performing and environmentally responsible. Several critical design factors can guide sustainable liquid cooling deployment:

• IT inlet fluid temperatures: Raising rack temperatures can unlock surprisingly large efficiency gains. In a simulated Paris data centre running at 40 kW per rack, a 20 ° C increase with non-adiabatic cooling cuts energy use by roughly 40 %, with most of the benefit realised in the first 5 ° C. For adiabatic cooling, the same temperature increase reduces energy by about 30 % while slashing water usage by 40-60 %. Beyond that, savings taper off and the local climate plays a key role in how much impact you see.

• Heat rejection: Heat rejection architecture is crucial. The most sustainable designs often use aircooled chillers with economiser modes, which rely on cool ambient air to reject heat and operate with near-zero water use. This stands in stark contrast to traditional cooling towers, which are water-intensive and carry a significant embodied carbon footprint from continuous water pumping and treatment.

• Component selection: Component selection and proactive maintenance play a major role in achieving sustainable operations. Prioritising high-efficiency components like pumps and heat exchangers minimises operational energy, while selecting high-quality parts with long lifespans reduces the embedded carbon associated with manufacturing and replacements. A coordinated control system, plus a proactive maintenance plan, keeps the system humming at peak efficiency.

• Heat reuse: Liquid-cooled data centres make heat reuse feasible and practical. In contrast, air-cooled facilities expel low-grade heat, making it inefficient and often impractical to repurpose. On the other hand, liquid cooling can expel higher-grade heat, opening the door to more useful applications such as district heating or industrial processes.

Other design choices can make a big difference for sustainable liquid cooling. Rack power density drives how efficiently heat is captured. CDU type shapes how effectively liquid reaches the racks. And coolant type influences both performance and environmental impact. Paying attention to these factors helps maximise efficiency while minimising the carbon footprint.

A blueprint for future-proofing AI infrastructure

Moving to liquid cooling requires careful planning and a forward-looking strategy. A practical blueprint includes:

• Plan in parallel: Physical infrastructure planning and IT planning must happen together. IT and facilities teams need to collaborate from day one to avoid costly situations where AI hardware sits idle while the infrastructure lags behind.

• Design for flexibility and scalability: Build designs that can handle multiple hardware generations. This often means hybrid setups that mix air and liquid cooling or deploying high-temperature chillers today to ease a future transition to directto-chip cooling and higher rack densities later.

• Partner early and often: Start with early collaboration among IT vendors, cooling specialists and system integrators. These partnerships bring crucial insights, help establish best practices and pave the way for a seamless, optimised system. The collaboration between Schneider Electric and NVIDIA on reference designs for specific hardware is a prime example of this ecosystem approach in action.

• Embrace sustainability as a core requirement: Tie cooling strategies to corporate ESG goals and regional regulations( think water restrictions and emissions targets) from the start. Liquid cooling ' s closed-loop design and lower energy profile make it a powerful tool for meeting these requirements.

Bottom line: Liquid cooling is critical for the AI era

In the age of AI, liquid cooling is a prime enabler of innovation and a cornerstone

Steven Carlini, Chief Advocate, Data Center and AI, Schneider Electric

of competitive advantage. The move to advanced liquid cooling is driven by the need to remove heat, yet it naturally brings sustainability gains that support corporate ESG objectives, including achieving energy and water efficiency, cutting carbon emissions, shrinking server footprints and supporting higher rack densities.

With the right data, strategic planning, flexible design and a thriving ecosystem of trusted partners, data centre operators can access the insights they need to make informed decisions on key factors to build the resilient, sustainable infrastructure that will power the AI-driven future. �

50 www. intelligentdatacentres. com