The ongoing maturation and adoption of Cloud computing and artificial intelligence (AI) are driving global demand for data centre capacity. Yet, while these tools have increased productivity, boosted creativity, and saved precious time and resources on mundane tasks, many have overlooked the enormous power and water requirements for cooling data facilities. This forms a conundrum for industry leaders who must design solutions for more responsible growth.
Digital demand creates a real-world thirst
The increasing appetite for data has come at a staggering cost. In 2023, data centres accounted for 560 billion litres of water use, a figure researchers expected to climb as AI expands.
The energy needs of AI systems’ supply chains are astronomical, especially in countries like the United Kingdom. Yet, the water footprint often remains hidden, extending well beyond facility walls.
Data centre water usage is an increasingly complex issue, underscored by a recently proposed UK operation. The intended Northumberland facility will be the country’s largest. Although it promotes a water-free direct cooling system, its indirect water footprint could be over 50 times higher than its direct use estimate. This equates to 124 million litres annually.
Embracing ground breaking cooling innovations
To address data centres’ extreme water demand, the global electronics industry is shifting away from unsustainable evaporative techniques toward more water-efficient innovations. These advanced solutions, such as direct-to-chip liquid cooling and waste-heat reuse, deliver superior thermal management with far less water.
The transformative power of liquid cooling
The transition from air to liquid cooling has been among the most crucial inventions within data centre infrastructure. A liquid coolant is more effective than air because water is 800 times denser and absorbs nearly 3,200 times as much heat per unit volume. The immense thermal conductivity is more readily available and captured through targeted, closed-loop systems.
Unlike conventional methods that cool an entire room, direct-to-chip or immersion technologies use a coolant to absorb heat directly from the central processing unit. The heated liquid circulates through a radiator for cooling before returning to the processor. This self-contained cycle is critical to reducing on-site water and energy consumption.
Capturing and reusing waste heat
The industry is exploring other ways to avoid overusing water, such as turning waste heat into a valuable resource. Data centre operators can capture and reuse thermal energy, reducing the water and electricity needed for cooling and generating a new source of revenue.
However, some caveats could prevent widespread implementation. As demonstrated by Germany’s recent legislation, it is difficult to succeed when targets fail to account for sophisticated infrastructure requirements. Existing heat networks are third-generation systems and incompatible with low-temperature heat from data centres. It could take over 10 years to fill a new facility, making the initial heat output uncertain.
Getting smarter about water sourcing
Increasing global water scarcity has created a dire need for smarter water sourcing for data facilities. According to a 2025 Global Digital Sustainability Alliance report, freshwater demand will likely outpace supply by 40% by 2030, suggesting that relying on municipal water is no longer justifiable.
This is a serious problem for the data industry, as 55% of the world’s data centres are in river basins with high water pollution risks. Another 68% of these facilities are close to protected areas and key biodiversity habitats.
Data centre operators must seek alternative sources to avoid placing additional stress on the water supplies in these regions. A viable strategy worth further exploration is treating and utilising captured rainwater, industrial runoff and grey water.
Charting a course for sustainable digital growth
Future digital growth hinges on a sustainable approach to cooling technological infrastructure. As data demand explodes, the environmental cost must not go unchecked. Closed-loop and waterless systems can be transformative as the industry aims to reduce its water footprint.
Engineering explainers
How to maximise battery life in wearable medical devices
As wearables become more common, more patients and providers expect incremental functionality, performance and longevity improvements.
How stream data processing enhances smart grid management
Global electronics has more information to process than it knows how to handle. Data scientists are engineering next-generation strategies to parse it all.
How to secure hardware against side-channel attacks
With the growing demand for faster, more connected devices, companies need stronger hardware security than ever. Encryption algorithms and secure protocols protect sensitive data, but lesser-known security threats – like side-channel attacks – can easily bypass these components.
Addressing 3 workforce challenges in electronics manufacturing
Electronics manufacturers in the UK face many workforce challenges also seen in other parts of the world. Tackling them is essential since the industry’s goods are in perpetually high demand. An awareness of the most pressing issues is also necessary before affected parties can begin mitigating the effects.
How to secure data in transit
Today, data is constantly moving, with people worldwide continuously sending emails, uploading files to the Cloud and sharing information between devices. This is known as data in transit, which is often vulnerable to attacks.
