AI and Water Usage
Gabriella Sandoval, Events Manager and Liaison
With technological advancements rapidly increasing, it is vital to recognize the negative impacts that these advancements pose. One particular area of technology that has transformed from a futuristic concept to mainstream technology is AI, whether that be ChatGPT or other AI services, that can generate any question or service in a split second.
According to Global AI statistics, as of 2025, around 378 million people use AI tools. Global AI has also reported that in just 2025, there have been nearly 65 million new users which is the highest annual increase yet. Also, one-third of new users have come from the U.S alone. Given these numbers, it’s no surprise that most AI users come from only a few countries with the majority from the U.S.
Detected disadvantages regarding the use of AI are privacy and security concerns, dependence and reliability issues, and the impact on human interaction. A concern that has been widely acknowledged is the use of water for cooling AI systems. The data centers operating AI servers use large amounts of electricity power. This generates an enormous amount of heat; and to keep servers from overheating, many facilities rely on these cooling systems by using water.
As data centers continue to multiply and consume vast amounts of resources, they’re creating new challenges for local communities. A prime example is in Newton County, Georgia, where one of Meta’s large data centers is located. Their massive water usage has caused water tables to drop, leaving some residents without access to water altogether. With no signs of slowing down, it’s uncertain what further consequences may take place if changes aren’t made soon. Large data centers have the ability to consume up to 5 million gallons per day. This is equivalent to the water use of a town populated by 10,000 to 50,000 people.
A 2024 report from the Lawrence Berkeley National Laboratory estimated that in 2023, U.S. data centers consumed 17 billion gallons (64 billion liters) of water directly through cooling. It projects that by 2028, those figures could double, or even quadruple. It is also estimated that by 2028, AI in the US could consume 300 terawatt-hours of energy annually. This is enough to provide electricity to over 28 million households; which is equivalent to as much as 720 billion gallons of water annually. The process of cooling AI servers costs the amount of water needed to meet the indoor needs of 18.5 million households through a year.
Due to the concerning global freshwater scarcity we are facing, increasing water use in data centers is unwise. According to the United Nations Environmental Report, nearly two-thirds of our world's population experiences severe water shortages for at least one month a year. By 2030, this gap is predicted to become much worse, with almost half of the world's population facing severe water stress.
As interest in AI keeps growing, data centers will likely keep expanding. But, without clear and consistent tracking of water use, neither the public nor the government will have the full picture in consideration. This impacts behavior when making decisions about where to build data centers, how to regulate them, and how they impact local communities and the environment.