AI's Hidden Thirst: UN Warns Data Centres Could Consume Water Needed by 1.3 Billion People by 2030

Key Takeaways

• UN researchers warn AI-driven data centres could double their water and electricity consumption by 2030.
• Water demand from AI infrastructure could reach levels equivalent to the annual needs of 1.3 billion people.
• AI workloads are expected to account for a growing share of global data-centre energy use.
• Governments are considering new efficiency standards and sustainability regulations.
• The long-term success of AI may depend on balancing technological innovation with environmental responsibility.

Artificial intelligence has become the defining technology race of the decade. Governments and technology giants are investing hundreds of billions of dollars into AI infrastructure, data centres, and advanced computing systems.

But a new report from researchers at the United Nations University Institute for Water, Environment and Health (UNU-INWEH) suggests that the rapid expansion of AI could create a major environmental challenge: water and electricity consumption. According to the report, AI-driven data centres are on track to double their use of power and water by 2030, potentially consuming enough water to meet the annual needs of approximately 1.3 billion people.

The findings add to a growing debate about whether the world's AI ambitions can be achieved sustainably.

Why AI Uses So Much Water

Most people associate artificial intelligence with software. However, every AI query, image generation request, or chatbot interaction relies on massive physical infrastructure.

At the centre of this infrastructure are data centres, large facilities packed with servers, processors, networking equipment, and storage systems. These facilities generate enormous amounts of heat and require extensive cooling systems to operate safely.

Water is often used in cooling processes to prevent servers from overheating. As AI models become larger and more computationally demanding, the cooling requirements increase significantly. Researchers warn that the growing deployment of generative AI systems is accelerating this trend worldwide.

Data Centres Could Double Resource Consumption by 2030

The UN-backed research estimates that data centres will consume roughly twice as much electricity and water by the end of the decade compared with current levels. The primary driver is the explosive growth in AI applications across industries, governments, and consumer services.

According to the report:

• Data centre electricity consumption could nearly double by 2030.
• Water demand for cooling systems is expected to rise sharply.
• AI workloads could account for around 40% of data-centre energy use by 2030, up from roughly 20% today.
• The environmental burden extends beyond energy to land use, cooling infrastructure, and electronic waste generation.

The report warns that without stronger oversight and sustainability measures, AI infrastructure could place unprecedented pressure on local resources.

The Water Challenge Is Bigger Than Data Centres Alone

Experts note that AI's water footprint extends beyond server cooling.

Water is also consumed indirectly through electricity generation and semiconductor manufacturing. Every AI chip requires water-intensive production processes, while power plants supplying energy to data centres often depend on large quantities of water for cooling.

This means the true water cost of AI may be significantly higher than what is visible at data-centre sites.

Some estimates suggest a single hyperscale data centre can consume hundreds of millions of litres of water annually, while a mid-sized facility may require more than a million litres each day.

Why Governments Are Paying Attention

The findings arrive as countries worldwide compete to become AI leaders.

Technology companies including Alphabet, Microsoft, Amazon, and Meta are investing heavily in new AI infrastructure. Industry spending on AI-related data centres and computing facilities is expected to reach hundreds of billions of dollars over the coming years.

However, growing concerns about water scarcity, electricity demand, and carbon emissions are pushing regulators to take action.

The European Union recently announced plans to develop stricter energy-efficiency standards and sustainability measures for data centres as power demand continues to rise.

Environmental groups and local communities have also increasingly questioned whether large AI facilities should be built in regions already facing water stress.

Can AI Become More Sustainable?

Technology companies argue that advances in cooling systems, renewable energy adoption, and infrastructure design can reduce AI's environmental footprint.

Several firms are investing in closed-loop cooling technologies, water-replenishment programs, and energy-efficient data-centre architectures. Google, for example, recently announced new commitments aimed at replenishing more water than it consumes by 2030.

Meanwhile, researchers are exploring alternative cooling methods, waste-heat recovery systems, and more efficient AI models that require less computing power.

Yet experts caution that efficiency improvements alone may not offset the sheer scale of expected AI growth.

The Bigger Question: Is AI Creating a New Resource Economy?

The UN report highlights a shift that is often overlooked in discussions about artificial intelligence.

The future of AI may depend not only on algorithms and computing power but also on access to water, electricity, land, and critical infrastructure.

As nations race to build larger AI systems, the debate is increasingly moving beyond technology and into resource management. The challenge for policymakers will be balancing innovation with sustainability before AI's environmental footprint grows faster than the solutions designed to contain it.