Today, it is difficult to find someone who does not use Artificial Intelligence (AI). From university assignments to office work, AI has become an everyday companion. While much has been written about its impact on jobs, mental health and misinformation, a far less discussed consequence is its growing environmental footprint, particularly its enormous demand for freshwater.
At first glance, the connection between AI and water may seem puzzling. After all, we interact with AI through smartphones and computers. Yet every question asked to an AI model is processed inside massive data centres filled with powerful servers and graphics processing units (GPUs). These machines generate enormous amounts of heat and must be cooled continuously to prevent failure.
Many AI data centres rely on evaporative cooling, where water absorbs heat as it evaporates, much like sweat cools the human body. As a result, AI consumes vast quantities of freshwater every day. In 2025, Google alone used around 10.9 billion gallons of water, a 34 per cent increase from the previous year and more than double its consumption in 2021. Globally, AI data centres consumed an estimated 560 billion litres (148 billion gallons) of water in 2025. A single large facility can use up to five million gallons daily, roughly matching the water needs of a city of 10,000 to 50,000 people.
The problem is not simply the volume of water consumed, but the type of water. Most cooling systems require freshwater because seawater corrodes equipment. Yet freshwater is one of Earth's scarcest resources. Although water covers nearly three-quarters of the planet, only about 3.0 per cent is freshwater, and only a tiny fraction is readily accessible in rivers, lakes and wetlands.
This growing demand comes at a time when around 2.2 billion people still lack safely managed drinking water. Worse still, many AI data centres are located in regions already facing water stress, placing additional pressure on local communities and agriculture.
AI is transforming the world, but its hidden environmental costs cannot be ignored. Technology companies must invest in water-efficient cooling systems and alternative technologies, while users should avoid unnecessary AI queries. The future of AI should not come at the expense of one of humanity's most precious and irreplaceable resources--- freshwater.
Laboni Akter Kobita
Student
Jagannath University
brunfelsiapauciflora7@gmail.com