The Hidden Thirst: A Growing Concern Over AI’s Water Footprint

In the race to develop and deploy advanced artificial intelligence, a hidden environmental cost is drawing increasing scrutiny: water consumption. As data centers, the “factories” of AI, expand at a rapid pace, their demand for massive amounts of electricity and, critically, water, is raising alarms among environmental advocates and policymakers. A recent Morgan Stanley report projects that AI data centers could increase their annual water use more than 11-fold by 2028, reaching over one trillion liters.

The core of the issue lies in the immense heat generated by AI servers, particularly during the energy-intensive process of training large language models. To prevent these servers from overheating, data centers rely on sophisticated cooling systems, many of which use evaporative cooling towers. This process, while effective, consumes a staggering amount of freshwater as it evaporates to dissipate heat. In fact, a study from the University of California, Riverside, and The Washington Post found that a single query to a generative AI model can require the equivalent of a small bottle of water.

The problem is compounded by the indirect water footprint of AI. This includes the water used to generate the electricity that powers the data centers, as well as the water-intensive process of manufacturing the semiconductor chips and other hardware that make up the AI infrastructure. More than half of the world’s top data center hubs are already located in regions facing medium to high water stress, creating localized impacts that can strain municipal water supplies and spark public protests.

Companies like Google and Microsoft have reported significant increases in their water usage, largely driven by their data center operations. For instance, Google’s Iowa facility alone reportedly consumed 3.8 billion liters of water in a single year. While some companies are exploring more efficient cooling methods, such as liquid and immersion cooling, and locating facilities in cooler climates, the overall demand continues to rise exponentially. The conversation around AI sustainability is shifting to address this overlooked crisis, emphasizing the need for greater transparency from tech companies and more efficient, water-conscious practices to ensure a viable future for both technology and the planet.