AI Water Use Tops 6.26 Trillion Gallons, Driven by Chips and Power as Demand Doubles by 2050
Updated
Updated · The Information · May 11
AI Water Use Tops 6.26 Trillion Gallons, Driven by Chips and Power as Demand Doubles by 2050
2 articles · Updated · The Information · May 11
6.26 trillion gallons of water were withdrawn for AI in 2025, and Xylem projects that total will more than double by 2050, with the biggest burden falling on chip factories and power plants rather than data centers.
50% to 70% cuts in freshwater use at hyperscale data centers have come from closed-loop cooling, but that shifts demand to electricity-intensive systems and the water used to generate that power.
40% of the world’s data centers already sit in high or extremely high water-stress areas, while 29% of chip factories are in extremely water-stressed regions, making location a key risk.
17 billion gallons used on site by U.S. data centers in 2023 remains small beside other sectors, and AI accounts for 3.7% of the industrial world’s 168.8 trillion gallons of annual water use.
30% of global water is lost through utility leaks and theft, prompting tech companies including Microsoft to fund AI-based leak detection in cities such as Phoenix and Las Vegas.
AI helps find water leaks, but its supply chain consumes trillions of gallons. Can technology solve the crisis it's creating?
Big Tech is building AI empires in water-starved lands. Is a global resource showdown with local communities inevitable?
The Soaring Water Footprint of AI: Sustainability, Scarcity, and the Race for Responsible Growth
Overview
The rapid expansion of Artificial Intelligence (AI) technologies is driving a significant increase in water consumption, raising urgent concerns about global resource management. This growing water footprint stems from the infrastructure needed to power and operate AI systems, including the manufacturing of advanced chips and the continuous operation of large data centers. These facilities require substantial energy, much of which is generated by water-intensive thermal power plants, and they consume large amounts of water for cooling. Without effective mitigation strategies, the current trajectory of AI development could lead to profound environmental consequences, making sustainability a critical challenge for the future of AI.