Artificial Intelligence’s Growing Water Footprint Amid a Global Water Crisis

Artificial Intelligence‘s Growing Water Footprint Amid a Global Water Crisis

PAHARI BARUAH

The explosive growth of artificial intelligence (AI) in 2025 has delivered unprecedented advancements in productivity, innovation, and daily life. Tools like OpenAI‘s ChatGPT and Google’s Gemini have become ubiquitous, processing billions of queries daily. However, this AI boom carries a substantial environmental cost, particularly in water consumption and carbon emissions.

Global water scarcity projections underscore regions where data center expansion risks intensifying local shortages.

A landmark study published in December 2025 in the journal Patterns by Alex de Vries-Gao, a researcher at Vrije Universiteit Amsterdam and founder of Digiconomist, provides the first comprehensive isolation of AI’s specific environmental impact.

The research estimates that AI systems alone could consume 312.5 to 764.6 billion liters of water in 2025-exceeding global annual bottled water consumption (approximately 446-468 billion liters). Simultaneously, AI’s carbon footprint could reach 32.6 to 79.7 million tonnes of CO₂ equivalent, comparable to the annual emissions of New York City (around 52 million tonnes).

These figures highlight how AI, powered by energy-intensive data centers, is straining freshwater resources at a time when over 2 billion people lack access to safe drinking water, and projections warn of severe scarcity affecting two-thirds of the global population by the late 2020s.

AI’s Dual Environmental Impacts: Carbon Emissions and Water Consumption

De Vries-Gao’s analysis, drawing on corporate disclosures and International Energy Agency (IEA) data, reveals AI’s 2025 carbon footprint rivaling that of major cities or sectors. The upper estimate equates to more than 8% of global aviation emissions. Critics, including Donald Campbell of UK non-profit Foxglove, argue that “the public is footing the environmental bill for some of the richest companies on Earth,” as tech giants reap profits while externalizing costs.

Water use is equally alarming. AI-related consumption surpasses prior estimates for all data centers combined by over a third, driven by both direct cooling and indirect electricity generation needs.

Charts illustrating the rapid growth in data center and AI-related water consumption.

Google Data Centers – Annual Water Consumption

(Units in millions for gallons and liters)

 The average Google data center consumed 550,000 gallons (2.1 million liters) of water per day, equivalent to 200 million gallons (760 million liters) of water annually.

How Data Centers and AI Consume Water

AI relies on hyperscale data centers filled with thousands of high-performance servers.

Direct Water Use

Evaporative cooling towers dominate, where water evaporates to dissipate heat-up to 80% of withdrawn water is lost. A hyperscale facility can consume millions of liters daily.

Indirect Water Use

Electricity generation, especially from thermal sources, accounts for two-thirds of the footprint. The IEA notes AI-focused centers rival aluminum smelters in power draw. The US (45% of global data center electricity), China (25%), and Europe (15%) dominate consumption. Many new facilities are sited in water-stressed areas for incentives or land availability. Examples include concerns in India over diesel backups creating “massive carbon liability” and UK projects emitting equivalent to thousands of homes.

The IEA projects total data center electricity doubling to 945 TWh by 2030, with AI as the primary driver-potentially pushing water use toward 1,200 billion liters annually.

De Vries-Gao criticizes insufficient disclosures: companies aggregate impacts, omitting AI-specific or indirect water use (e.g., Google’s Gemini reporting). He calls for stricter transparency to fairly allocate costs.

Pathways to Mitigation

Innovations include immersion cooling, closed-loop systems (reducing freshwater use by 70%), air cooling, and recycled water. Companies like Microsoft aim for “zero-water” designs; pledges for water positivity by 2030 exist, though usage has risen. Strategic siting in cooler climates or renewable grids, plus policy mandates, are crucial.AI holds promise for solving water challenges-through leak detection, irrigation optimization, and scarcity modeling. Yet its unchecked growth risks aggravating the crisis.

As de Vries-Gao questions: If tech companies benefit most, should society bear the costs? Sustainable AI demands transparency, innovation, and shared responsibility to ensure technology serves humanity without depleting vital resources.

Mahabahu.com is an Online Magazine with collection of premium Assamese and English articles and posts with cultural base and modern thinking.  You can send your articles to editor@mahabahu.com / editor@mahabahoo.com (For Assamese article, Unicode font is necessary) Images from different sources.