Not All Data Center Technologies Cause Water Crises
Clean water is a basic necessity for society, while technological investment should contribute to improving public welfare rather than creating new competition that undermines local domestic needs. Nevertheless, one issue requires closer examination: are data centers actually a direct cause of water shortages, as suggested by media coverage of Batam, or do these problems reflect a combination of drought, infrastructure limitations, growing industrial demand, and weaknesses in water resource governance?
This question matters because debates over data center water consumption have emerged in various countries, including the United States. Concerns about water use by the AI industry are not entirely unfounded. Several studies and government reports have identified genuine risks. However, these risks cannot be generalized to all data center facilities. The scale of their impact depends heavily on cooling technology, water sources, facility capacity, climatic conditions, and the environmental carrying capacity of their operational locations.
Water and Digital Infrastructure: Real but Uneven Risks
Data centers are essential infrastructure for the modern digital economy. They store, process, and distribute information that supports a wide range of services, from financial transactions and cloud computing to digital communications and AI development. Growing computational demand, particularly for large-scale AI models, is driving the construction of data centers with increasingly high electricity and cooling requirements.
Servers generate substantial heat during computational operations. To maintain operational stability and prevent equipment damage, data centers require cooling systems. Some systems use water through evaporative processes, while others rely on closed-loop liquid circulation, air-based cooling, or combinations of different technologies.
This is where water consumption becomes an issue. Evaporative cooling systems can consume significant volumes of freshwater because some of the water evaporates and does not return to its original source. Other cooling technologies can reduce direct water consumption, although they may require greater energy input or additional infrastructure.
Research by Lei and colleagues (2025), published in Resources, Conservation and Recycling, indicates that water use across data center workloads can vary by more than 10,000 times. This variation is influenced by server efficiency, water consumption in electricity generation, equipment utilization rates, cooling systems, and climatic conditions. The finding is important because it demonstrates that not all data centers have the same water footprint or consumption patterns.
A study by Privette, Barros, and Cai (2026), published in AGU Advances, emphasizes that aggregate data center water use may appear relatively small compared with total national water consumption. However, local impacts can become significant when facilities are concentrated in regions already experiencing water stress. The researchers also identify limited transparency in corporate water-use data as a major obstacle to effective policymaking and public oversight.
In other words, the central issue is not simply how much water data centers consume, but where that water comes from, how its use is managed, and whether such consumption reduces water availability for local communities.
The United States: When AI Expansion Meets Water Resource Constraints
Debates over data center water consumption have also intensified in the United States, one of the world’s leading centers of AI industry growth. Several regions, including Virginia, Arizona, Texas, and Georgia, are facing questions about whether their water and electricity infrastructure can adequately support data center expansion.
In Virginia, home to one of the world’s largest concentrations of data centers, a study of groundwater conditions in the Eastern Coastal Plain warned that local aquifers are already under pressure and could face further depletion due to population growth and industrial demand. Virginia has hundreds of operational and planned data center facilities. However, the report did not conclude that data centers are the sole cause of groundwater depletion. Most facilities rely on public water systems supplied primarily by surface water, while comprehensive data on the industry’s groundwater withdrawals remain limited.
These findings demonstrate that the central problem does not always lie in proven current consumption, but also in the government’s capacity to measure future demand before approving new industrial developments.
The case of Georgia presents a different picture. In Fayetteville, concerns over water use by a data center campus owned by Quality Technology Services (QTS) emerged after water authorities discovered previously unrecorded water usage. An E&E News report in May 2026 stated that more than 29 million gallons of water had not been properly recorded in the billing system. The company subsequently explained that the water had been used for construction activities rather than operational cooling and stated that its facility employed a closed-loop cooling system.
This case does not prove that data centers directly caused water shortages among local residents. However, it highlights the importance of industrial water-use records, connection transparency, and public utility oversight.
