Can Data Centres Keep Getting Bigger Without Getting Thirstier?

The answer may depend on how intelligently we design the systems around the servers.

Cooling efficiency cannot be measured only by how much heat a system can remove. Data centre operators also need to consider how much electricity and fresh water that cooling process requires. This makes water management an important part of the overall MEP strategy rather than a separate sustainability initiative.

Smarter controls can make a difference here. Sensors and building management systems can continuously monitor temperature, humidity, pressure and cooling demand. This allows cooling equipment to respond to actual conditions instead of operating at a fixed capacity throughout the day. Better controls can reduce unnecessary cooling and help operators identify inefficiencies before they become expensive problems.

The location of a data centre matters too. Climate, water availability and local infrastructure should influence the cooling strategy from the earliest stages of design. A solution that works well in a water-abundant region may not make sense in an area facing water stress. Designing for local conditions can reduce the environmental burden while improving long-term resilience.

There is also a broader opportunity in the relationship between cooling and energy. When heat is captured at a higher temperature through liquid cooling, it can become easier to consider heat recovery for suitable applications. Instead of rejecting all that heat into the atmosphere, facilities can explore ways to reuse it for building heating, industrial processes or other nearby applications where the temperature and economics make sense.

This points towards a different way of thinking about data centre growth. The goal should not simply be to build larger facilities. It should be to build facilities that deliver more computing capacity per unit of energy, water and physical infrastructure.

Liquid cooling can play an important role in that transition. Direct-to-chip systems, immersion cooling and secondary fluid networks can support the higher rack densities required by AI and high-performance computing while reducing the dependence on conventional air-based cooling. Comfonomics, for example, works across liquid cooling and MEP infrastructure for high-density data centre environments.

The industry is already moving in this direction. As AI workloads become more demanding, cooling infrastructure is becoming a core part of data centre architecture rather than something added around the computing equipment. Comfonomics has also expanded its capabilities in this area through Secondary Fluid Network infrastructure designed for high-density liquid-cooled environments.

“The bigger question is no longer whether data centres will grow. They almost certainly will.

The real question is whether their resource consumption has to grow at the same rate.”

With better thermal engineering, closed-loop systems, intelligent controls, efficient MEP design and greater attention to water reuse, the industry has an opportunity to decouple digital growth from rising resource consumption.

That is where Comfonomics comes in. Through its focus on engineering, efficiency and sustainable infrastructure, Comfonomics is working towards data centres that can support the next generation of computing without treating water and energy as unlimited resources.