How AI Uses Our Clean Water
The rapid advancement of artificial intelligence (AI) is increasingly becoming a part of daily life, but behind its convenience lies a rarely discussed environmental issue: water consumption. A BBC World Service report highlights that the more than one billion messages sent to AI chatbots each day require substantial resources, including water.
AI systems depend on data centres, which house high-powered servers and computer chips. When a user sends a command, the system performs complex calculations, generating significant heat. To prevent damage, data centres must be continuously cooled. One common method is liquid cooling, where a coolant absorbs heat from the chips and is then cooled using water. In some processes, up to 80 percent of this water can evaporate.
The water used is typically clean, potable water, the same resource needed for communities, agriculture, and sanitation. This has raised concerns about the pressure data centres place on local water supplies. Protests against data centre construction have occurred in countries including Spain, India, Chile, Uruguay, and parts of the United States, driven by fears over water and electricity grid impacts.
Beyond direct cooling, AI also consumes water indirectly through its massive electricity usage. Data centres require enormous amounts of power, and the thermal or nuclear power plants that generate this electricity also use significant volumes of water in their operations. The International Energy Agency predicts that global data centre electricity demand could more than double by 2030, reaching approximately 945 terawatt-hours, with AI-optimised centres being a primary driver.
Some major technology companies are beginning to address the issue. Microsoft has pledged to replenish more water than it consumes by 2030, while Google and Meta have announced similar commitments focusing on efficiency, water restoration, and transparency. However, these efforts are not yet sufficient to counterbalance the rapid expansion of AI. More robust oversight, innovation in water-efficient cooling, use of renewable energy, and siting data centres with local environmental conditions in mind are crucial steps to mitigate the negative impacts. The digital progress promised by AI must be balanced against its reliance on natural resources, requiring cooperation between society, governments, and tech companies to ensure a sustainable future without sacrificing clean water availability.