Data centres are becoming larger and more powerful. They are the engines driving digital evolution, but they face two major problems. The first is well known: they consume an enormous amount of energy. The second, much less explored but potentially even more dramatic: they require a gigantic volume of water.

The Water Consumption of AI
Artificial intelligence, a technology demanding in terms of computing capacity, exacerbates the problem. According to a study by the Universities of Arlington and Riverside (USA), global water withdrawal for AI-related data centres could reach between 4.2 and 6.6 billion cubic metres by 2027. To put that in perspective, it's roughly half of the UK's annual water consumption and five times that of Denmark.

These figures represent a real concern at a time when, according to a report published by UNESCO on behalf of UN-Water, 26% of the global population suffers from a lack of safe drinking water and 46% lack access to safe sanitation services. Furthermore, according to UNESCO, global water demand could increase by 30% in the next 25 years. While the water withdrawal of AI-related data centres by 2027 is expected to be between 4.2 and 6.6 billion m³.

Possible Solutions
Data centres mainly use water for cooling purposes. The machines that process and store information generate heat and require temperature control to function optimally. Various cooling technologies exist, but the choice depends on local climatic conditions.

To mitigate environmental impact, data centres can draw on alternative water sources such as collected rainwater, seawater, or treated wastewater. These solutions fit within ESG (Environmental, Social, and Governance) frameworks and Water Usage Effectiveness (WUE) standards, which are particularly relevant in regions at risk of water scarcity. Some advanced systems recycle water or reuse it for local communities through sophisticated seawater cooling systems.

Investing to Accelerate Change
Some companies are already making great strides, supported by financial investments that can enable the ecological transition to accelerate. Asset managers, in fact, have the ability to direct capital towards the most promising companies, supporting them while also creating medium-to-long-term financial opportunities for investors.

“We believe that emerging technologies,” explains Michael Stewart, Head of Pooled Index Strategy in the Asset Management division of L&G (Legal & General, a major British financial services company), “are set to play a significant role in reducing water consumption and improving the ESG performance of data centres.”

Kao Data, for example, is one of the companies that develops and manages data centres in which L&G has invested. It is among the signatories of the Climate Neutral Data Centre Pact, an agreement aimed at reaching net-zero water consumption in data centres by 2030.

Towards Smart Water Management
If technological evolution (and artificial intelligence) brings new challenges, it also offers possible solutions.

Take the case of water networks: predictive AI, by analysing huge amounts of data, can identify and report areas where infrastructure is most degraded, suggesting targeted interventions before significant failures or leaks occur. Efficiency also involves water distribution. Modern cities, with increasingly complex networks, must ensure a steady flow without waste. This is where algorithms come into play again, capable of analysing data in real time, optimising water flow and pressure in pipes. The result? Greater reliability and a more resilient network, even during maintenance or repairs.

But the revolution is not just about quantity, because water quality is another crucial aspect. Advanced sensors can monitor in real time key parameters such as pH (a measure of acidity or alkalinity), turbidity, and the presence of contaminants. This means fewer risks, immediate interventions, and greater safety for those who consume drinking water.

Artificial intelligence can be decisive not only on a local scale but also globally, guiding water resource management strategies. In a world increasingly exposed to droughts and growing water demand, the ability to analyse availability levels and consumption trends in real time allows governments and companies to make more informed decisions, promoting fairer and more sustainable distribution.

The integration of artificial intelligence into water treatment systems is the final piece of a revolution taking shape. Real-time monitoring, remote management, process optimisation – all contribute to making infrastructures more efficient, sustainable, and secure.

“As well as being essential to our daily activities,” Stewart continues, “the infrastructures that support digitalisation are fundamental to achieving key social goals, including greater financial inclusion and environmental sustainability.”

See:L’intelligenza artificiale ha sete