Artificial intelligence's explosive growth is creating a hidden environmental crisis that extends far beyond carbon emissions: a massive demand for water that could strain global supplies and pit technology companies against farmers and communities already facing scarcity. A comprehensive United Nations University report published this week warns that the rapid expansion of AI infrastructure is driving growing environmental pressures through its increasing demands for electricity, water, and land, with researchers calling for urgent action to ensure the technology develops within planetary limits.

How Much Water Does AI Actually Need?

The numbers are staggering. Data centers consumed an estimated 448 terawatt-hours of electricity in 2025, with AI-related workloads accounting for about 20 percent of that demand. But electricity is only part of the story. The associated water footprint of data centers could reach 9.3 trillion liters annually by 2030, enough to meet the drinking water needs of the world's population for approximately 1.6 years. To put that in perspective, separate research by water technology company Xylem and Global Water Intelligence projects that water demand across the entire AI value chain could increase by 129 percent by 2030, adding around 30 trillion liters of annual water demand.

These projections matter because water scarcity is already a critical issue in many regions. As AI infrastructure expands into arid or drought-prone areas, competition for water between technology firms, agriculture, and communities becomes a potential flashpoint. For investors and policymakers, this creates several layers of risk that go beyond environmental concerns.

Where Is AI Infrastructure Creating the Most Pressure?

The geographic concentration of AI infrastructure amplifies these risks. Only 32 countries host AI-specialized cloud infrastructure, and around 90 percent of global AI computing capacity is concentrated in the United States and China. This concentration means that certain regions will bear disproportionate environmental costs while others reap the economic benefits.

According to water experts, the greatest future pressures are likely to emerge in specific sectors and regions. Christopher Gasson, chief executive of Global Water Intelligence, noted that the pressure points are concentrated in semiconductor manufacturing and in fast-growing data center hubs in the United States, East Asia, and South Asia. These are precisely the regions where water stress is already a concern, making the situation particularly urgent.

What Are the Key Environmental and Social Risks?

The environmental footprint of AI extends far beyond carbon emissions and includes significant impacts on water resources, land use, mineral extraction, and electronic waste. For sustainable investors and policymakers, understanding these interconnected risks is essential to managing the technology's long-term impact.

• Water Scarcity Risk: Datacenters require large volumes of water for cooling, often in regions already facing water stress. As AI infrastructure expands, competition for water between technology firms, agriculture, and communities becomes a potential flashpoint that could disrupt operations and erode companies' social license to operate.
• Regulatory and Physical Risk: Municipalities may impose water-use restrictions or pricing reforms in response to datacentre expansion, while physical scarcity could disrupt datacentre operations themselves. Investors in datacentre operators, cloud service providers, and enabling infrastructure must now assess water exposure with the same seriousness historically reserved for mining or agriculture.
• Electronic Waste Generation: AI infrastructure could generate up to 2.5 million metric tonnes of electronic waste each year by the end of the decade, creating disposal and recycling challenges, particularly in developing nations that often bear the burden of waste management.
• Inequality and Geopolitical Risk: Many countries bear the environmental costs associated with mineral extraction and electronic waste disposal while only a handful benefit from AI computing capacity, deepening global inequalities and creating potential policy backlash.

"The future of artificial intelligence should not be measured only by what machines can do, but by whether humanity can deploy those capabilities within planetary boundaries. Though often described as weightless and virtual, the reality of AI is profoundly physical. Behind every prompt, image, or video lies a growing infrastructure of energy systems, water withdrawals, land use, mineral extraction, and electronic waste," said Professor Kaveh Madani, director of the United Nations University Institute for Water, Environment and Health.

Professor Kaveh Madani, Director, United Nations University Institute for Water, Environment and Health

Can Technology Solutions Address the Water Problem?

Despite the dire projections, experts argue that the challenge is not insurmountable. The report emphasizes that the environmental footprint of AI is not fixed; it is shaped by infrastructure, energy sources, model design, how much AI is used, what it is used for, and where that use takes place. This means that decisions made now by governments, companies, researchers, and users will determine the scale of AI's future environmental footprint.

Matthew Pine, president and CEO of Xylem, emphasized that many of the tools needed to address the challenge already exist. Advanced treatment technologies allow water to be recycled rather than wasted, and digital systems can help manage supply in real time, reducing water lost to leaks. The key is targeted investment and collaboration between industry, utilities, and governments.

"AI is placing new demands on water supplies, but many of the tools needed to address the challenge already exist. Advanced treatment technologies, for example, allow us to recycle water rather than waste it. Digital systems can help better manage supply in real time, reducing water lost to leaks. It's time for a water transition built on targeted investment and collaboration between industry, utilities, and governments to ensure water systems can support both growth and community resilience," said Matthew Pine.

Matthew Pine, President and CEO, Xylem

What Steps Can Stakeholders Take to Manage AI's Water Impact?

• Expand Wastewater Reuse Programs: Investment in wastewater reuse, leakage reduction, and digital water management systems could offset much of the anticipated growth in water demand. In regions facing the greatest pressure, expanded wastewater reuse and leakage reduction can fully offset future growth in AI-related water consumption.
• Implement Transparent Lifecycle Accountability: Responsible AI development requires greater transparency, international cooperation, and lifecycle accountability. Companies and governments must track and disclose the full environmental cost of AI infrastructure across the entire supply chain, from chip fabrication to datacentre operations.
• Shift Behavioral Patterns and User Choices: Every prompt, default setting, generated image, video, and query accumulates when multiplied by billions of users and thousands of operators worldwide. Behavior change across this entire decision chain, from individual users to corporate planners, is one of the most powerful and underused levers for keeping AI within planetary limits.
• Integrate Water Risk into Investment Frameworks: Investors should assess water exposure in datacentre operators and cloud service providers with the same rigor applied to mining or agriculture companies. Portfolio construction, engagement priorities, and risk assessment frameworks must evolve to account for AI's water footprint alongside carbon considerations.

The UN report concludes that decisions made by governments, technology companies, investors, infrastructure operators, and users will determine whether AI's benefits can be achieved without placing unsustainable pressures on global environmental systems. As AI electricity demand is expected to double or triple by 2028 from 2024 levels in the United States alone, the window for action is narrowing. The technology itself is neither inherently sustainable nor unsustainable; its net effect emerges from the interaction between markets, policy, and capital allocation.