Ecolab: The Mutually Rising Tides of Data and Water

As AI and real‑time analytics become business‑critical, data centres globally are under pressure to scale compute density – and with it, cooling capacity. 

Yet amid the focus on megawatts and Power Usage Effectiveness (PUE), a quieter crisis is unfolding: the growing impact of this computing power on our already-stressed freshwater resources. 

Data centre water consumption is not solely from on-site evaporative cooling. On average, 4.3 litres of water are consumed to generate a kWh of energy. 

With data centre power use projected to double by 2030 in Europe, water stress will only increase. 

Here, we speak with Matt McMullen, Senior Corporate Account Manager, Global High Tech at Ecolab, on the mutually rising tides of data and water.

Why does water matter in data centre cooling?

“Cooling towers and other evaporative systems can provide energy efficiency, reliability, and cost savings and have long been the backbone of facility‑scale temperature control,” begins Matt.

“However, in regions classified as water‑stressed, even municipal supply can face seasonal shortages. A key industry improvement to deal with this challenge is the use of data-driven decision making, namely Water Usage Effectiveness (WUE).” 

WUE, first introduced by The Green Grid in 2011, considers litres of water used per kWh of IT load to provide operators and data centre design professionals with a comparable KPI, much as PUE did for power. 

“Added analysis accounting for water consumed from power generation (WUE), local water stress factors (WS) and use of sustainable water sources have helped give a more accurate KPI for actual local water stress impact,” says Matt. 

“These WUE metrics show that innovations like adiabatic cooling and the use of greywater, or industrial reclaim, have helped relieve some water stress. 

“Another potentially large WUE improvement has come as a result of leveraging direct-to-chip (D2C) liquid cooling to answer the AI-driven need to cool chips with higher heat densities.”

How is direct‑to‑chip cooling a differentiator in sustainability?

“Direct‑to‑chip (D2C) liquid cooling uses coolant flowing through cold plates mounted directly on processors to remove heat 100 times more efficiently than air alone,” explains Matt. “This technology also permits higher return water temperatures, enabling partial free cooling on cooler days and nights, decreasing the need for additional chiller energy. 

The result has a dual benefit: improved energy efficiency and reduced water consumption related to power generation and evaporative cooling. 

“By circulating water-based coolant at rack level, we reduce demand on the chillers and cut evaporative losses,” explains Michael Obradovitch, Global Area Vice President for Data Centres at Ecolab. “This not only lowers WUE but also brings water use decision-making right down to where the heat is generated.” 

In fact, life cycle assessment studies suggest advanced cooling technologies are “reducing greenhouse gas emissions (15–21%), energy demand (15–20%) and blue water consumption (31–52%) in data centres.”

How can data centre operators establish standards and best practices in light of the current challenge?

“The promise of D2C comes with a caveat,” notes Matt. “Small variations in coolant chemistry, even the more forgiving but less energy-efficient, glycol-based coolants, can lead to corrosion, fouling or microbial growth, reducing capacity, increasing PUE and jeopardising uptime. That’s why pre‑commissioning cleanliness is a non‑negotiable.”

Ecolab co‑authored the Open Compute Project’s (OCP) new guide on pre‑commissioning Thermal Control System manifolds, outlining hydro‑flushing, particulate removal and final water‑quality checks to protect sub‑100 m channels.

‘Design Clean → Maintain Clean’ is the mantra of Mukul Girotra, Senior Vice President and General Manager of Ecolab’s Global High-Tech Division. 

“By integrating real‑time monitoring from day one, operators can catch drift in pH, conductivity, or glycol mix before it escalates, ensuring both uptime performance and sustainability,” says Mukul.

Real-time monitoring networks feed central dashboards with leading indicators of fluid health, empowering engineers to preempt issues rather than react to system failures.

What is the future of sustainable cooling?

“Looking ahead, edge AI sites and micro data centres, from London’s financial district to remote 5G hubs in Dubai, will all demand even more water‑smart designs,” says Matt. 

“Predictive analytics, driven by AI itself, will forecast chemistry drift and automate operational adjustments.”

The industry’s vision, according to Matt, is net‑zero water impact through circular‑economy cooling loops, where reclaimed water and captured condensate feed directly back into closed‑loop systems.

“In this future, every drop, like every watt, will be accounted for and optimised,” concludes Matt.