Human society’s demand for data and information in support of technology is accompanied by a large demand on environmental resources. Today’s Data Center Facilities are warehouse structures that house networked computers to provide computing necessary to meet demands imposed by Artificial Intelligence (AI), streaming, cloud storage, and other technological advances. Data Centers are proliferating worldwide with the US leading the way with 5,426 facilities in 2025 or ~50% of all Data Centers. The individual facilities are also increasing in size and capacity. In 2010 most data processing facilities existed within the institution that they supported. By 2020 these internal facilities were matched in number with off-site stand-alone centers. Presently, newer AI Data Center ‘Campuses’ with over 1 Gigawatt capacity are becoming more commonplace. The facilities so far in Pennsylvania number >40 with several others proposed. An interactive map showing existing and proposed Data Center locations can be found here.
The magnitude of computing power and speed at these facilities translates into a large demand for energy and water for maintaining and cooling the processors. In 2023, Data Centers globally used 4.4% of total energy consumption, which is expected to increase to ~10% by 2028, equivalent to energy use by multiple large cities. On-site water consumption rates in 2023 were 70 billion liters annually, approximately 3X the consumption in 2019. The on-site water consumption for Data Centers is not necessarily higher on a per area basis relative to other industries such as retail centers or hotels. It is the comparatively large area of Data Centers, especially hyperscale Data Center complexes, that increases the total demand for water. Further, the on-site consumption is lower than the indirect demand for water because water is required by the utilities to provide for the energy consumption by Data Centers.
To provide some further perspective, water withdrawal directly associated with Data Centers represents <0.2% of total water withdrawal in the US. There is, however, the opportunity for withdrawal to negatively impact local water supplies and waterways. Locally, water utilities are strained to meet public demand amidst challenges associated with unpredictable supply associated with droughts and floods and uncertain regulations being imposed by emergent contaminants such as PFAFs and microplastics. Therefore, action at the local level can be critical in helping to alleviate these stressors on the water industry that are compounded by the recent increase in Data Centers. In addition to the strain on local water and energy resources, the large physical footprint of the warehouses often lead to loss of wildlife habitat and stormwater management issues.
The rapid demand for Data Centers is outpacing adequate time for decision-making and planning around their location and support. Locally and at the state level, there is work on legislation aimed at easing the demands for energy and water by Data Centers including placing the burden of energy and water support directly on the facilities. Other advances such as closed water supply systems, air-cooled facilities, free-air cooling systems, and on-site stormwater retention ponds can also help increase efficiency and/or decrease demand for water by Data Centers. Another example is underground Data Centers such as those in Norway that can also help to decrease demand for cooling by taking advantage of the relatively stable belowground temperatures similar to passive geothermal cooling systems.
We can all play a role in helping us move forward intelligently with new Data Center proposals. Residents concerned with plans for new or existing facilities can reach out to their municipalities and encourage proper scrutiny with how the facility will shift the availability of local groundwater and surface water supplies and with energy requirements. Municipalities can work with their utility providers to provide adequate planning. Individuals can also help with demand. By taking a break occasionally from AI and other high-energy technologies, and using the computing power of the human brain, we can potentially reduce demand, slow the pace of the current facility expansion, and permit more time for proper planning.
Sources:
Map shows growing number of data centers planned across Pennsylvania
Data Center Mapper
Data Centers and Water Use in Pennsylvania
Data Center Water Usage: A Comprehensive Guide
Siddik, Shehabi, and Marston. 2021. The Environmental Footprint of Data Centers in the United States. Environmental Research Letters 16 064017.
L. Marston. 2025. Pennsylvania State University Water Cooler Talk