As operators evaluate how best to future-proof their facilities, many are taking a closer look at where their energy is actually going—and where the biggest gains can be made. Cooling remains one of the most resource-intensive parts of any data center, making it a natural starting point for efficiency improvements. The shift toward liquid cooling is accelerating in part because it aligns with broader goals: lowering operating costs, reducing environmental impact and supporting the high-density compute that AI and modern workloads demand.
How Liquid Cooling Works
Liquid cooling typically relies on two primary approaches: direct-to-chip cooling, where coolant absorbs heat directly from processors, and immersion cooling, where hardware is fully submerged in a thermally efficient dielectric fluid. Both methods offer significantly higher heat-removal capacity than air cooling, making them well-suited for dense, AI-driven compute environments.
Driving Better Power Efficiency
Power Usage Effectiveness (PUE) is the industry standard metric for measuring data center energy efficiency. It compares the total power consumed by the facility to the power used specifically by IT equipment. It’s defined as:
PUE = Total facility energy ÷ IT equipment energy
An ideal PUE, representing perfect efficiency, is 1.0. Forbes suggests real-world data centers fall some way short of this, with current average PUE values standing at around 1.55 globally. This average has been gradually improving over the years, but the overall figure masks considerable variability.
Air-cooled centers typically achieve a PUE of 1.5–1.7, while those in liquid-cooled facilities can reach 1.2, with immersion cooling capable of pushing this even lower. According to DataGarda, the most efficient hyperscale data centers achieve PUEs as low as 1.06.
In fact, using PUE as the only measure of cooling effectiveness may lead one to underestimate the benefits of liquid cooling. Because it can reduce the energy consumed by both IT equipment and the cooling system, it can shrink both elements of the equation shown above. Some commentators believe an alternative measure of Total Usage Effectiveness (TUE) would be a better way of measuring these benefits.
Efficiency Gains and Operational Considerations
The real-world impact is significant. Data from Boyd shows that shifting from 100% air cooling to a 75% liquid-cooled environment cuts cooling facility power use by 27% and lowers total site power by more than 15%. These savings illustrate how liquid cooling delivers efficiency gains beyond what PUE alone captures—driving benefits across both the cooling system and the IT load it supports.
While retrofitting liquid cooling into legacy spaces may require structural or plumbing adjustments, many operators find that the long-term efficiency gains outweigh the initial complexity.
Reducing Water Consumption and Environmental Impact
As well as the efficiency gains, liquid cooling also performs well against other sustainability measures, including reduced water consumption since it removes the need for evaporative cooling towers. BlueField Research notes that water use varies significantly between traditional evaporative systems and modern closed-loop designs:
- Evaporative cooling systems on average consume <0.5 gallons (<1.9 liters) per kWh, with less efficient installations consuming >0.5 gallons (>1.9 liters)
- Closed-loop liquid cooling systems use <0.03 gallons (0.1 liters) per kWh, since they continuously recirculate the same water
Since most data centers rely on municipal water supplies, access to water remains a major constraint on facility location and growth. Less water-intensive closed-loop liquid-cooling systems place a reduced burden on public utilities, enhancing their appeal for new developments and potentially reducing hurdles from local authorities and surrounding communities.
By minimizing evaporation, scaling and biological growth risks, closed-loop systems deliver both operational and environmental benefits without compromising reliability.
A Strategic Shift Toward Sustainable Operations
As operators look ahead, liquid cooling is emerging as a strategic path to lower resource use, support higher-density AI workloads and meet evolving sustainability expectations. By reducing both energy and water consumption while improving overall system stability, it gives data centers a more resilient foundation for future growth—enabling smarter, cleaner and more efficient operations in an increasingly demanding digital landscape. When paired with ongoing monitoring of flow, temperature and water quality, liquid cooling helps operators sustain performance and extend system longevity over time.
