By Canninah Dladla, Cluster President for Sub-Saharan Africa at Schneider Electric
As AI workloads push data centre rack densities to unprecedented levels, cooling efficiency is becoming as important as the ability to remove heat.
There are seven core design factors and three operational practices that dictate energy and water consumption.
DC Studio | Magnific.com
For data centre and HVAC engineers, the challenge is no longer simply to provide sufficient cooling capacity, but to balance energy consumption, water use, operating temperatures and the changing demands of high-density computing.
The sustainability of direct-to-chip liquid cooling therefore requires a holistic approach extending from the chip through the cooling distribution system and ultimately to the chiller and heat-rejection system.
The logic of efficiency – the 7+3 Framework
Sustainability in the AI era is governed by the trade-offs between energy and water consumption. Optimising Power Usage Effectiveness (PUE) requires a sophisticated understanding of approach temperatures (the temperature difference between the primary and secondary cooling loops).
The sustainability and efficiency of a direct-to-chip liquid cooling system are driven by a holistic view of the architecture, from the chip to the chiller.
There are seven core design factors and three operational practices that dictate energy and water consumption.
Seven design factors:
- Outdoor Heat Rejection: Air-cooled chillers with economizers save water versus towers, which consume heavily year-round.
- IT Inlet Fluid Temps: Raising supply temps (e.g., 25°C → 45°C) cuts compressor energy 30–40% and water use 40–60%.
- Rack Power Density: Higher AI-driven densities boost liquid heat capture, slightly improving plant efficiency.
- CDU Type: Liquid-to-Liquid CDUs are more efficient than Liquid-to-Air, avoiding extra heat exchange cycles.
- Component Selection: Efficient pumps/heat exchangers with low approach temps enable higher facility water temps, more economiser hours, and less compressor energy.
- Coolant Type: PG25 is standard for reliability; deionised water offers better thermal capacity but requires complex chemical management.
- Heat Re-use: Higher liquid cooling temps enable waste-heat recovery for district heating/industry, improving Energy Reuse Factor.
Three operational practices:
- Control Scheme: Advanced systems use digital twins and dynamic pump control to match real-time loads, avoiding wasted energy.
- Maintenance: Regular coolant monitoring and filter cleaning prevent damage and reduce pump strain.
- Adaptive Deployment: Modular, phased builds align infrastructure with evolving IT demand, avoiding stranded capacity and excess carbon.
Designing for AI is an exercise in mastering fluid dynamics and material science at scale. Moving away from anecdotal risks toward a physics-based architectural strategy will enable facility engineers to deploy direct-to-chip cooling with confidence.
