Understanding water use in data centers requires contextualized systems thinking

This water use is especially pressing in the Asia-Pacific region, where the data center market is expanding at an extraordinary speed. (Image: iStock)

As artificial intelligence and cloud computing continue to reshape industries, the infrastructure powering this transformation— data centers —is facing growing scrutiny. While energy consumption has long dominated sustainability conversations, water use is now emerging as a critical concern.

This issue is especially pressing in the Asia-Pacific region, where the data center market is expanding at an extraordinary speed. Recent analysis from Cushman & Wakefield shows that the APAC development pipeline is outpacing North America, solidifying its position as the world’s fastest-growing region for digital infrastructure. With hyperscale facilities rising across countries such as India, Indonesia, and Vietnam—often in water-stressed zones—concerns around water consumption and long-term resilience are becoming central to discussions.

Headlines have warned that data centers in Melbourne’s north and west could consume enough drinking water to supply 330,000 residents annually. Globally, AI-driven computing is projected to withdraw up to 1.7 trillion gallons of water by 2027, more than four times Denmark’s annual water use.

These figures are alarming. But are they a cause for real concern?

My answer is: it depends.

The complexity behind water use

Water use in data centers is highly variable. The impact depends on a range of factors—technology, location, climate, and energy source. Some facilities use air-cooled systems that consume little to no water. Others rely on water for cooling, using evaporative cooling towers or adiabatic air handling, which are more efficient in hot climates but require significant water input.

Cooling demand also fluctuates. In many cases, water is only used during peak summer conditions when air cooling alone is insufficient and when communities most need potable water supplies. Annual water use may be far lower than peak-day figures suggest.

Some water consumption figures quoted for the data center sector include off-site water use, such as “Scope 2” water consumed in generating electricity, or “Scope 3” water consumed in the manufacture and construction of the data center.  Electricity generation—especially from fossil fuels—can involve substantial water withdrawals, although not always from the same drinking water supplies. Renewable sources like wind and solar, by contrast, use virtually none.

This leads us to a critical insight: water and energy are deeply interconnected. Decisions made to optimize one resource can inadvertently strain another. For example, prioritising energy efficiency might lead to water-cooled systems in water-scarce regions. Conversely, minimising water use could increase energy consumption and carbon emissions.

The Water-energy nexus: Why systems thinking matters

This interdependence between water and energy is known as the water-energy nexus. It’s a concept that demands systems thinking—an approach that considers the whole lifecycle and environmental context of infrastructure decisions.

data centers consume electricity, which generates heat, requiring cooling—often with water. The more energy consumed, the more cooling is needed. But the choice of cooling system affects both water and energy efficiency. The optimal solution depends on local climate, water availability, energy mix, and operational priorities.

This is why blanket regulations, such as uniform water usage efficiency (WUE) or power usage efficiency (PUE) targets, can be counterproductive. Without considering local conditions, they risk incentivising the wrong behaviours. For example, a narrow focus on energy efficiency may prompt operators to adopt water-cooled systems in water-stressed regions.

Instead, we must not evaluate data centers in isolation, but as part of a broader environmental and infrastructure ecosystem.

Each project requires tailored design

There is no one-size-fits-all solution when it comes to water use in data centers. Each project must be assessed within its specific environmental and infrastructural context. The local climate plays a significant role. Cooler regions may allow for more efficient air-cooled systems that use minimal water, while hotter climates might necessitate water-intensive cooling technologies.

The availability of potable water is another critical factor; using drinking water for industrial cooling in water-stressed areas raises ethical and sustainability concerns. Additionally, opportunities for water reuse or alternative sources—such as recycled water, harvested rainwater, or industrial byproducts—can significantly reduce the net impact on local water systems.

A facility designed with recycled water in a water-abundant region may have minimal environmental impact. However, applying the same design in a drought-prone area could exacerbate local water stress and lead to community or regulatory pushback. This is why regulatory frameworks must reflect this complexity. Blanket mandates that ignore local realities risk driving suboptimal decisions and unintended consequences.

In this context, systems thinking becomes not just useful, but essential. It enables decision-makers to balance trade-offs, optimize performance, and minimise environmental impact by considering the full spectrum of interrelated factors—water, energy, climate, and community needs.

Water and energy are deeply interconnected. Decisions made to optimize one resource can inadvertently strain another. (Image: iStock)

Opportunities to improve water sustainability

Despite the challenges, there are meaningful ways to reduce water impact and improve sustainability:

Technology innovation

Emerging technologies offer real promise. Chip-level cooling, advanced water treatment systems, and hybrid cooling solutions can reduce both water and energy use. Innovation must be encouraged, not stifled, by regulation.

Circular water thinking

Facilities can be integrated into the local water cycle. This includes using recycled wastewater or harvested rainwater, reusing cooling tower blowdown, and exploring waste heat recovery. Circularity isn’t just about reducing waste—it’s about creating value from what was once discarded.

Localized design

Cooling systems should be tailored to climate and water availability. Nature-based solutions can help manage stormwater, reduce urban heat, and enhance biodiversity. Climate resilience should be built into every design.

Active operational mindset

As a relatively cheap input cost, the importance of monitoring and improving water efficiency performance is often lower down the priority list.  Regular efforts to understand water system performance and vulnerabilities can improve operational reliability and provide insights for future designs and upgrades. 

Strategic location

With fast data transport, workloads can be shifted to regions with abundant renewable energy, cooler climates, or sustainable water supplies. Location strategy is a powerful lever for environmental performance.

These strategies require collaboration between engineers, sustainability teams, regulators, and communities. They also require transparency—clear reporting of water use, energy consumption, and environmental impact.

Embracing systems thinking

Some media narratives have overstated the water threat posed by data centers. In reality, they represent a small share of global water use. But as a fast-growing industry central to our digital future, data centers must lead by example.

The real issue isn’t whether data centers use water—it’s how thoughtfully they do so. A systems approach that balances energy, water, climate, and community impact is not just good practice—it’s essential for long-term resilience.

Embracing systems thinking requires close collaboration between stakeholders across the value chain. We need more informed dialogue, more transparency, and more integrated thinking. Only then can we ensure that the infrastructure powering our digital lives is not just fast and reliable, but sustainable.

---

This article is republished in collaboration with Ramboll. Written by Antony Gibson, director, water technology at Ramboll