https://orcid.org/0000-0002-1309-0039
Figures
Abstract
Datacenters are powering the Artificial Intelligence (AI) revolution. However, their water insecurity risks remain neglected. Limited research on the matter quantifies water demand at national or watershed-scales and estimates water use associated with training and using AI models. Research fails to examine water insecurity concerns held by households and communities where datacenters are planned or are operational. This article identifies four water insecurity concerns in the U.S. by synthesizing public reporting and legal filings involving non-governmental organizations, citizen coalition groups, investigative reporters, and individual citizens. These concerns include how datacenters’ development and operation can (i) undermine the democratization of water governance; (ii) contribute to unsustainable water use and rising utility costs; (iii) reduce the flexibility and resilience of water use decision-making; and (iv) increase water use across scales as a result of rising electricity demand. Three areas for future research are identified from the cases reviewed. First, local governments and utilities do not always readily provide water use data associated with datacenter operations; hence, public records should be requested and shared to democratize decision-making. Second, water-related risks posed to public health, rural and land-based livelihoods, and ecosystems from datacenter operations require context-specific empirical investigation. Third, examining whether and how specific water governance arrangements can engender acute health, economic, and environmental risks, especially under extreme events such as heatwaves or droughts, requires institutional analyses. Overall, analyzing datacenters’ volumetric water use within local contexts offers a more relevant analysis of water insecurity concerns and experiences.
Citation: Shah SH (2026) Four water insecurity concerns about datacenters driving the AI revolution. PLOS Water 5(1): e0000500. https://doi.org/10.1371/journal.pwat.0000500
Editor: Caitlyn Hall, The University of Arizona, UNITED STATES OF AMERICA
Published: January 13, 2026
Copyright: © 2026 Sameer H. Shah. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Partial support for this research came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant, P2C HD042828, to the Center for Studies in Demography & Ecology at the University of Washington. The content is solely the responsibility of the author and does not necessarily represent the official views of the National Institutes of Health. Additional support for this research came from support to CSDE from the College of Arts & Sciences, the UW Provost, eSciences Institute, the Evans School of Public Policy & Governance, College of Built Environment, School of Public Health, the Foster School of Business, and the School of Social Work. Further, in his role as a John C. Garcia Term Professor, Dr. Shah recognizes the generous financial support made possible by Carole Garcia. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The author has declared that no competing interests exist.
Introduction
Clean, affordable, and reliable water supports food production, electricity generation, industrial activities, and basic household functions [1]. However, in the United States (U.S.), researchers and communities have challenged assumptions that water access is near-universal, trustworthy, clean, and affordable [2–4]. Existing water access challenges are likely to be compounded by future freshwater use, with projections ranging from as low as an 8% decrease in use to as high as a 235% increase in the next fifty years depending on socio-economic, environmental, and infrastructure changes [5]. Some even argue that infrastructure decay, institutional rigidity, and weather and climate change now place the U.S. in a post-peak water security period whereby “[…] more and more households [will] manage with unsafe, inadequate, expensive, or unreliable water” [6]. This context can enable “datacenters”—the computing infrastructure backbone powering the Artificial Intelligence (AI) revolution—to accelerate existing water insecurity risks or contribute new ones.
Datacenters are facilities that store and operate computing infrastructure essential to the innumerable digital services now integral to daily life. The AI revolution will require unprecedented investments in datacenter infrastructure and, in particular, “hyperscalar” datacenters (HDCs). HDCs are vast, high-performance facilities engineered for scalability to manage the intense demands of data processing, distribution, and storage for generative AI, digital content creation, cryptocurrency mining, and other computing tasks [7–9]. A recent analysis estimated that global “datacenter demand” could easily triple by 2030 and, in the U.S., surge by 20–25% annually during this period [10]. The same study reported that HDC development will drive around 70% of the 2030 worldwide datacenter demand [10].
The operation of datacenters and HDCs often require large volumes of water to be withdrawn and consumed both directly and indirectly. Withdrawals refer to the total appropriation of water and consumption refers to the volume of water used by datacenters that is now no longer available for other human and ecological uses [11]. Direct consumption cools infrastructure [12]. Researchers estimated that datacenters directly consumed 21.2 billion liters of water in the U.S. in 2014 [13]. This volume tripled to about 66 billion liters of water in 2023 [13]. In 2028, U.S. HDCs alone could directly consume between 60–124 billion liters of water [13], equivalent to the annual use of around 530,000—1.1 million Americans. In contrast, indirect water use refers to water withdrawn or consumed for electricity generation, both to power datacenters and to treat their direct water use [12]. Indirect water uses can eclipse direct uses of water. In 2023, U.S. datacenters consumed around 176 TWh, which equates to about 800 billion liters of water, after accounting for regional differences in fuel source mixes [13].
The impacts of datacenters on water resources have almost exclusively assessed or forecasted direct and indirect volumetric water use associated with their operations [13,14]. In other cases, the potential to create or exacerbate water stress is identified based on the spatial location of datacenters [12]. Research has also estimated water use associated with training and using AI models [14,15]. To the author’s best knowledge, there are no existing studies about how datacenters and, in particular HDCs, affect local “water security”—a construct that reflects the degree to which households and communities can continue accessing, using, and benefiting from existing water systems [16]. Household water insecurity is not only associated with short-term water supply interruptions or long-term declines [17]. Water insecurity manifests through difficulties in food preparation, challenges in maintaining hygiene, and in the feelings of exclusion and citizenship related to decision-making and governance [17–21]. Community-scale water insecurity can influence household-level water insecurities and, vice-versa, household-level water insecurities can reveal communal-scale risks. Community-level data captures water insecurity dimensions such as the capacity for hazard mitigation and the future stability of water rates that are not necessarily knowable by relying only on household-level metrics.
This article presents four water insecurity concerns at household and community-scales associated with datacenter development and operation in the U.S. context. It recognizes that the scientific community and the public at-large lack adequate information of water insecurity concerns associated with datacenters’ direct and indirect water use [14]. The concerns identified reflect areas of risk identified by non-governmental organizations, citizen coalition groups, investigative reporters, and individual citizens in U.S. legal filings and public reporting. This article urges researchers to move beyond abstract quantifications of volumetric water use towards analyzing the potential experiential consequences of this for communities and ecosystems.
Concern #1: Undermining the democratization of water governance
“Non-disclosure” refers to the withholding of information concerning direct water use and/or the water-related risks for consumers or ecosystems associated with datacenter operation. More broadly, it also involves the selective presentation of data, such as not collecting or reporting indirect water requirements. For example, the use of more efficient closed-loop “non-evaporative” cooling systems are laudable innovations for improving direct water use efficiencies. Conveying these efficiencies without reporting the amount of water used to power datacenters can be misleading. In the U.S. context, the “indirect” water footprint of datacenters is ~ 12-times that of the volume of “direct” water use at the national-scale [13].
Both forms of non-disclosure are incompatible with the principles of democratic water governance. Democratic water governance is a collaborative and deliberative process of designing, implementing, and monitoring water management actions for a wide range of situational and place-based challenges [22]. It allows for different water users to shape the strategic nature of water infrastructure investments, provisioning, and use decisions in ways that serve their needs and wants—yet this process is only possible if data and information is publicly available. Acts of non-disclosure involve different actors, including datacenter clients (e.g., corporations), city and county governments, and water utilities. Concerningly, court filings from across the U.S. already reveal how these actors leverage public record exemptions to avoid water use data disclosure. This has included characterizing data as “confidential” information [23], company “trade secrets,” [24,25], and even as revealing of “personal financial information” [26].
In 2024, a case was filed against Dorchester County, South Carolina by a concerned citizen on allegations of a violation of the state’s Freedom of Information Act resulting from the county not disclosing water use data for Google’s $510 million “Project Orchid” datacenter [24]. Dorchester County and Google claimed, “data and information related to [Google’s] actual or projected consumption or usage of all or any portion of, or the amount of, the [water or sewer needs] shall be deemed Confidential Information in accordance with S.C. Code §30-4-40(a)(1)” [24]. S.C. Code § 30–4-40(a)(1) exempts “trade secrets,” or the “unpatented, secret, commercially valuable plans, appliances, formulas, or processes” from public disclosure [27]. The case was settled in 2025 and Dorchester County agreed to disclose Google’s water use data [28,29]. Significant delays in producing public record requests about datacenter water use have also been a basis for filing legal action. In September 2025, the Milwaukee Riverkeeper, Inc. filed a lawsuit because its request for the water consumption figures for Microsoft’s large, 1,575-acre datacenter campus under construction in Racine County, Wisconsin was not reasonably met under state public records law [23]. The filing, alleging an “unlawful delay in producing records,” came 210 days after the Milwaukee Riverkeeper, Inc. submitted the original public records request to the Racine Water Utility [23]. The Wisconsin Department of Justice states a reasonable time to respond to a “simple” request is 10 business days [23]. Shortly after filing, the data was released, indicating the campus could use as much as 234,000 gallons per day (GPD) on a “peak day,” and ~ 2.8 million gallons per year (GPY) by 2026 [30]. This figure could increase by three-times to ~ 8.4 million GPY at a future date as later phases of the project are completed [30]. Finally, the City of Colorado Springs, in creative fashion, declined a similar public records request on the basis that the release of water use data could constitute a disclosure of “personal financial information” and allow the “habits of individuals” to be identified [26]. Disclosing the gallon amount of water use could, the city indicated, be combined with the rate charge for the datacenter customer to reveal financial information about them [26]. The Court rejected this argument, writing the price per unit of water paid “[…] is simply the cost of water usage, not personal financial information” and required the disclosure of water use data [26]. These examples are concerning because they indicate an unwillingness for local governments and utilities to share data on public resources that affect the communities they represent and serve. Moreover, non-disclosure of this nature can undermine the informed participation of local stakeholders and rightsholders in water planning processes.
Concern #2: Contributing to unsustainable water use and rising utility costs
Datacenters, especially HDCs, risk creating or exacerbating unsustainable and inequitable water-related risks at local and regional scales. Examining consumptive, on-site water uses absent of the experiential, place-based risks such appropriations engender does not yield meaningful impact-based analysis. Legal filings and investigative reporting reveal concerns related to hydrological sustainability and rising utility costs. Two examples illustrate such concerns.
In late 2024, members of the Coalition for Responsible Data Center Development sued the City of Farmington, Minnesota, seeking a temporary injunction over a proposed $5 billion datacenter so that potential hydrological, economic, environmental, and health impacts could be better understood [31]. The Coalition alleged, in part, the datacenter could double the city’s water demand from 2.14 to 4.49 million GPD, and that the city “failed to properly evaluate the significant water resource implications of these developments” by not conducting assessments about “well siting, aquifer sustainability, and contingency plans for water shortages” [31]. The filing identifies water-related risks raised by concerned residents, who comprise the Coalition. These include the contamination of “high vulnerability” Drinking Water Supply Management Areas, the potential for “groundwater depletion in interconnected aquifers” and declines in stream water for trout, and the “[p]otential contamination of private wells [...] which will require residents to dig deeper wells at the homeowners’ expense” [31]. This lawsuit, now consolidated under Castle Rock Township v. City of Farmington et al. (2025) [32], does not only identify the amount of water projected to be used by the datacenter, but the concerns from it that encompass sustainability, emergency preparedness planning, and homeowner costs.
In Newton County, Georgia, Meta (formerly: Facebook) has datacenter operations in the Stanton Springs Industrial Park. Meta’s datacenter operation comes under two entities named in recent documents as, “Morning Hornet LLC,” and “Baymare LLC,” with the latter having “almost double the capacities for water and wastewater service agreements” [33]. These operations exist alongside a major pharmaceutical company (Takeda) and an under-construction Rivian electric vehicle construction plant. The Industrial Park, including Meta’s operations, relies on water provided by the Newton County Water & Sewer Authority (NCWSA). As of 2025, Meta is one of the “top ten customers” of the NCWSA in terms of annual water usage [33]. Their operations currently use 196,342 GPD, or around 70.7 million GPY [33]. However, by the county’s own estimate, that figure is expected to skyrocket by 2030 to between 390,000–1 million GPD, owing to growing water use from both facilities [34]. Due in part to Meta’s operations and other major industrial operations, the NCWSA reports a “potential water supply deficit in 2035” [34]. A recent New York Times [35] investigation reported, “If the local water authority cannot upgrade its facilities, residents could be forced to ration water. In the next two years, water rates are set to increase 33 percent, more than the typical 2 percent annual increases.” Official data from NCWSA shows the average residential water-sewer bill (5,000 gallons per month) from 2015—2022 remained almost constant ($72.57—$73.94/month) but, by 2025, increased to $85.10/month [33,36]. Water-sewer bills are projected to increase to $111.08/month in 2029 [33]. This projection would amount to an over 50% increase in the monthly water-sewer bill for NCWSA customers since 2015 (non-inflation adjusted), with the 6% average annual increase from 2023—2029. The NCWSA recognizes “uncertainty in the housing market and high inflation rates” have impacted purchases and operations and, importantly, disclose that “[c]ontinued infrastructure investment is necessary to meet residential and commercial growth,” explicitly citing the second Meta datacenter campus (Baymare), which “continues to give impetus to maintaining the capital improvement program in high gear” [33]. The cases from the City of Farmington and Newton County raise concerns for how water allocations could alter place-based human and environmental relationships with water and impact local and community life.
Concern #3: Reducing the flexibility and resilience of water use decision-making
As it is clear, datacenters can withdraw and consume large volumes of water. The potential risks associated with significant water allocations partly depends on the nature of the water use contract, local and regional environmental contexts, energy source, and future local infrastructure planning [37]. Rigid, fixed water use allocations for datacenters can undermine the flexibility of water managers to manage acute “shocks” to water users, shift the burden of adaptation onto them (e.g., water restrictions), and even impact the viability of future urban and rural development opportunities. Beyond carrying the risk of constraining decision-making, risk mitigation, and new opportunities, emerging evidence suggests datacenter development could increase the water insecurity of local communities to such an extent where water deficits may occur. This places a burden on water managers, local ecosystems, and potentially public citizens who—depending on how and whether datacenter operators contribute to infrastructure expansion—could share the cost. One example serves to illustrate this concern beyond the evidence cited for Newton County above.
In The City of The Dalles, Oregon—a city of only 15,884 (2024)—Google has operated datacenters since 2006. In 2021, the city denied a public records request submitted by a local news organization (The Oregonian) seeking disclosure of the annual amount of water the city provided to Design, LLC (i.e., Google), an industrial water user. Like other cases above, the city claimed Google’s water use was a company “trade secret” under Oregon state law (ORS § 192.345(2)) and was not subject to public disclosure [25]. The Oregonian subsequently sought review of this denial by submitting a petition to the Wasco County District Attorney’s Office, with that office ultimately issuing an order compelling the city to disclose records responsive to the request. While the city filed a lawsuit challenging the District Attorney’s Office decision, the parties ultimately reached a settlement agreement in December 2022 regarding disclosure of annual water usage. As part of the agreement, the city agreed to disclose annual water use data for 2012–2021 as well as for future years where publicly requested [38]. The data reveals that since 2012, Google’s water demand has increased by about 342% in The City of the Dalles (Fig 1). In 2024, Google’s local water demand was 461.1 million GPY (1.26 million GPD)—30% of the average daily demand for the city of the Dalles (2024: 4.2 million GPD) [39,40]. Looking to the future, the “average daily demand” (ADD) for the city is estimated to increase by 40.5% by 2029 relative to 2024 [40]. This demand increase appears almost entirely attributable to a 61.5% projected increase in non-residential demand over the same time period [40]. By 2034, the total city-wide ADD is expected to have increased by around 76.2% relative to 2024 levels. Over this period, residential demand is expected to remain nearly constant (increasing from 1.6 million GPD (ADD) to 1.7 million GPD) whereas non-residential users are projected to increase their demand from 2.6 million GPD (2024) to 5.7 million GPD—a 119% increase [40]. The city now faces a potential water deficit of 90 million gallons by 2034 [40].
Concern #4: Increasing water use across scales through significant energy demand
It is tempting to consider datacenters as discrete and bounded infrastructure facilities. In reality, datacenter development and operation relies on political priorities, regulatory contexts, and infrastructure systems near and far away from their physical location. This is compatible with how social scientists have reconceptualized infrastructure, “[…] not as individual objects but as parts of geographically spread socio-technical configurations: configurations which involve many different technologies, relations, capacities and operations, entailing different risks and power relationships” [44]. For instance, datacenters can simultaneously rely on decentralized water infrastructure operated by small water districts for direct cooling, even while having interconnections to a large, state-wide power grid. Thus, water insecurities can emerge near and far through the heterogenous “configurations” that enable datacenters to operate. One water insecurity concern involves the production of electricity, which can be both on-site and drawn from grid-connected generation facilities, and dwarf the water required for on-site cooling.
The “Stargate Project” serves as an example. Stargate is a multinational joint venture between OpenAI, Oracle, and SoftBank that will commit $500 billion USD to AI infrastructure in the U.S. over the next four years. “Stargate 1,” a $1.1 billion USD datacenter campus that is part of the larger “Stargate” venture, is under construction in Abilene, Texas. Stargate 1 reportedly plans on using an efficient closed-loop “non-evaporative” cooling system. Crusoe Energy Systems, LLC, a vertically integrated company leading the development of Stargate 1, reports “[t]he initial fill for the closed-loop system in each of the eight buildings requires one million gallons of water, sourced from the municipal water supply” [45]. An additional 101,000 GPY is estimated for maintenance purposes (i.e., 12,625 gallons per building each year) [45]. This direct water use is significantly lower as compared to other datacenters, including those smaller in scale identified herein. The campus has, however, contracted a 1.2 GW interconnection [46]. In addition, plans to construct a large 360 MW gas plant to supplement electricity requirements was revealed through a permit application [47]. The on-site gas plant could emit approximately 1.6 million tons of CO2eq per year; however, no estimates for annual water consumption are publicly available [47]. Moreover, no estimates exist for water use indirectly used for electricity production received through the 1.2 GW interconnection. The Houston Advanced Research Center (HARC) estimates that “data center water intensity” in Texas, or the average amount of water required for each MWh used, was 793 gallons per MWh [48]. Assuming a constant water intensity is required, a datacenter with Stargate 1’s electricity demands could require as much as ~ 8.3 billion GPY. The water insecurity risks associated with the enormous energy consumption required by datacenters now and in the near future will depend on the fuel-mix or source of power chosen. The HARC [48] estimates that meeting datacenter energy loads using natural gas with coupled carbon capture technology could use about 50- and 1000-times more water than solar and wind energy, respectively. Overall, water insecurity risks exist near and far from the physical site of a datacenter, and are related to different infrastructures, regulatory contexts, and strategic planning of those places.
A call for future research
Researchers who are concerned with the water insecurity risks posed by datacenters should not limit their analysis to quantifying how much water, either directly or indirectly, datacenters withdraw or consume. Exploring the volumetric appropriation of water within local and place-based social, political, and environmental contexts will offer a more complete and less abstract analysis of water insecurity concerns. Taking the context of the U.S., this article demonstrates these concerns through surveying a range of legal filings and court decisions, corporate sustainability reports, investigative reporting, and city and utility planning documents. The concerns raised by individuals and civil society organizations include water affordability, water sustainability, corporatization of water resources, compounding effects of multiple datacenters, short and long-term water deficits, contamination, environmental flows, and the barriers to public records.
From this exercise, I outline three future research inquiries needed to better understand the potential water insecurity risks associated with datacenters. First, local governments and utilities do not always readily provide water use data for datacenters. In fact, the legal cases above reveal long delays exist in responding to public record requests and demonstrate how exemptions for public records are leveraged to protect water use data. Relatedly, datacenter operators may establish secondary companies with names that are not easily traced back to the central company. This can undermine basic transparency and shield companies from public exposure, even without the intent to do so. Given evidence of non-disclosure, the most immediate role researchers can play is in requesting and sharing public records in the interest of democratizing water use decision-making. Second, it is difficult to determine when and under what conditions datacenters will receive water from municipal or other sources. Most legal filings surveyed here request disclosures of water volume. Researchers should investigate whether and how specific water governance arrangements can engender acute health, economic, and environmental risks, including how users are prioritized in the water allocation process during emergencies such as heatwaves or droughts. Third, very little, if any, attention has focused on non-drinking water impacts associated with datacenter development and operation. Given water is central to public health, rural and land-based livelihoods, and ecosystems, it is vital to understand risks and concerns posed to these uses. Overall, analyzing datacenters’ volumetric water use within local contexts offers a more relevant analysis of water insecurity concerns and experiences.
Acknowledgments
The author extends gratitude to C.L. Workman, L. Wright, and the participants in the Center for Studies in Demography and Ecology’s Grant Writing Summer Program for constructive comments provided on this research. The author dedicates this article to the late K. Horton.
References
- 1. D’Odorico P, Davis KF, Rosa L, Carr JA, Chiarelli D, Dell’Angelo J, et al. The Global Food‐Energy‐Water Nexus. Reviews of Geophysics. 2018;56(3):456–531.
- 2. Meehan K, Jepson W, Harris LM, Wutich A, Beresford M, Fencl A, et al. Exposing the myths of household water insecurity in the global north: A critical review. WIREs Water. 2020;7(6).
- 3. Méndez-Barrientos LE, Fencl AL, Workman CL, Shah SH. Race, citizenship, and belonging in the pursuit of water and climate justice in California. Environment and Planning E: Nature and Space. 2022;6(3):1614–35.
- 4. Wilson NJ, Montoya T, Arseneault R, Curley A. Governing water insecurity: navigating indigenous water rights and regulatory politics in settler colonial states. Water International. 2021;46(6):783–801.
- 5. Warziniack T, Arabi M, Brown TC, Froemke P, Ghosh R, Rasmussen S, et al. Projections of Freshwater Use in the United States Under Climate Change. Earth’s Future. 2022;10(2).
- 6. Jepson W, Wutich A, Pearson AL, Beresford M, Brewis A, Cooperman A, et al. Beyond peak water security: Household-scale experiential metrics can offer new perspectives on contemporary water challenges in the United States. PLOS Water. 2025;4(8):e0000413.
- 7.
Ermakov A. Electricity demand growth for data centres and AI and implications for natural gas-fired power generation. Gas Exporting Countries Forum (GECF). 2024.
- 8.
Lin L, Wijayawardana R, Rao V, Nguyen H, Gnibga EW, Chien AA. Exploding AI Power Use: an Opportunity to Rethink Grid Planning and Management. In: The 15th ACM International Conference on Future and Sustainable Energy Systems, 2024. 434–41. https://doi.org/10.1145/3632775.3661959
- 9. Kachris C, Z. Patrikakis C. The Rise of Accelerator-Based Data Centers: Opportunities and Challenges. IT Prof. 2024;26(6):4–9.
- 10.
McKinsey & Company. The data center balance: How US states can navigate the opportunities and challenges. McKinsey & Company. 2025. https://www.mckinsey.com/industries/public-sector/our-insights/the-data-center-balance-how-us-states-can-navigate-the-opportunities-and-challenges
- 11. Mytton D. Data centre water consumption. npj Clean Water. 2021;4(1).
- 12. Siddik MAB, Shehabi A, Marston L. The environmental footprint of data centers in the United States. Environ Res Lett. 2021;16(6):064017.
- 13.
Shehabi A, Smith SJ, Hubbard A, Newkirk A, Lei N, Siddik MAB. 2024 United States Data Center Energy Usage Report. Berkeley, California: Energy Analysis and Environmental Impacts Division, Lawrence Berkeley National Laboratory. 2024.
- 14. Herrera M, Xie X, Menapace A, Zanfei A, Brentan BM. Sustainable AI infrastructure: A scenario-based forecast of water footprint under uncertainty. Journal of Cleaner Production. 2025;526:146528.
- 15. Lo L. AI has a hidden water cost − here’s how to calculate yours. The Conversation. 2025.
- 16. Jepson W, Budds J, Eichelberger L, Harris L, Norman E, O’Reilly K, et al. Advancing human capabilities for water security: A relational approach. Water Security. 2017;1:46–52.
- 17. Young SL, Boateng GO, Jamaluddine Z, Miller JD, Frongillo EA, Neilands TB, et al. The Household Water InSecurity Experiences (HWISE) Scale: development and validation of a household water insecurity measure for low-income and middle-income countries. BMJ Glob Health. 2019;4(5):e001750. pmid:31637027
- 18. Pearson AL, Jepson W, Brewis A, Osborne-Gowey J, Wutich A, Beresford M, et al. A protocol for the development of a validated scale of household water insecurity in the United States: HWISE-USA. PLOS One. 2025;20(8):e0330087. pmid:40788878
- 19. Wutich A, Brewis A, Tsai A. Water and mental health. WIREs Water. 2020;7(5).
- 20. Wutich A, Ragsdale K. Water insecurity and emotional distress: coping with supply, access, and seasonal variability of water in a Bolivian squatter settlement. Soc Sci Med. 2008;67(12):2116–25. pmid:18954928
- 21.
Acevedo-Guerrero T, Bossenbroek L, Leonardelli I, Zwarteveen M, Kulkarni S. Routledge Handbook of Gender and Water Governance. Routledge. 2024. https://doi.org/10.4324/9781003100379
- 22. Koebele EA, Méndez‐Barrientos LE, Nadeau N, Gerlak AK. Beyond engagement: Enhancing equity in collaborative water governance. WIREs Water. 2023;11(2).
- 23. Milwaukee Riverkeeper. Milwaukee Riverkeeper v. City of Racine, Racine Water Utility, and Anjuman Islam. 2025.
- 24. Frank Heindel v. Dorchester County. 2024.
- 25. City of the Dalles v. Michael Rogoway and Advance Local Media LLC. 2021.
- 26. City of Colorado Springs, Beckler H. Businessinsider.com. 2025.
- 27.
South Carolina State Legislature. South Carolina Code of Laws - Title 30 (Public Records) Chapter 4 (Freedom of Information Act). https://www.scstatehouse.gov/code/t30c004.php
- 28. Frank Heindel v. Dorchester County. 2025.
- 29. Wren D. SC county to disclose Google data center’s water usage to settle legal battle. Post and Courier. 2025.
- 30.
Wisconsin Public Radio. Microsoft data centers will use up to 8.4M gallons of water each year, records show. https://www.wpr.org/wp-content/uploads/2025/09/MKE-Regional-Demands-030124.pdf. 2025.
- 31. Coalition for Responsible Data Center Development, Drea Doffing, Brian Haskin, Gary Johnson, Cathy Johnson, Mark Pearson, Terrie Pearson, Catherine Peregrino, and Jeff Schettler v. City of Farmington. 2025.
- 32. Consolidation - Coalition for Responsible Data Center Development, Drea Doffing, Brian Haskin, Gary Johnson, Cathy Johnson, Mark Pearson, Terrie Pearson, Catherine Peregrino, and Jeff Schettler v. City of Farmington. 2025.
- 33.
Newton County Water and Sewerage Authority NCWSA. FYE 2025 Budget and Financial Plan. Covington, Georgia: Newton County Water and Sewerage Authority. 2025. https://ncwsa.us/wp-content/uploads/NCWSA-%20Budget-Financial-Plan-2025.pdf
- 34.
Newton County, Newton County Water & Sewerage Authority, City of Covington. One Water Resources Analysis. Covington, Georgia: Newton County Water & Sewerage Authority. 2024. https://ncwsa.us/wp-content/uploads/One-Water-Resources-Analysis-2024-1.pdf
- 35. Their water taps ran dry when Meta built next door. The New York Times. 2025.
- 36.
Newton County Water and Sewerage Authority NCWSA. FYE 2022 Budget and Financial Plan. Covington, Georgia: Newton County Water and Sewerage Authority. 2022. https://ncwsa.us/wp-content/uploads/NCWSA-2022-Budget-FinalMH.pdf
- 37. Lei N, Lu J, Shehabi A, Masanet E. The water use of data center workloads: A review and assessment of key determinants. Resources, Conservation and Recycling. 2025;219:108310.
- 38.
City of the Dalles v. Michael Rogoway and Advance Local Media LLC. 2022.
- 39.
Environment Report 2025. Mountain View, California: Google. 2025. https://sustainability.google/reports/google-2025-environmental-report/
- 40.
Water System Master Plan Update: November 2024. The City of the Dalles. 2024. https://cms8.revize.com/revize/dallesor/departments/public%20works/master%20plans/cotd-water-system-master-plan-update-nov-2024-final.pdf?t=202502191842100&t=202502191842100
- 41. Rogoway M. Google’s water use is soaring in The Dalles, records show, with two more data centers to come. The Oregonian/OregonLive. 2022.
- 42.
Sessel A. Is Google a bad neighbor? A fight over water use at a huge datacenter is exposing deeper issues in an Oregon town. https://fortune.com/longform/google-data-center-the-dalles-oregon-water-dispute/. 2023.
- 43.
Environment Report 2024. Mountain View, California: Google. 2024. https://sustainability.google/reports/google-2024-environmental-report/
- 44. Lawhon M, Nilsson D, Silver J, Ernstson H, Lwasa S. Thinking through heterogeneous infrastructure configurations. Urban Studies. 2017;55(4):720–32.
- 45.
Crusoe Inc. An inside look at the Abilene AI data center. https://www.crusoe.ai/resources/blog/an-inside-look-at-the-abilene-ai-data-center. 2025.
- 46.
Lancium. Lancium and the Stargate Project in Abilene, TX: Bringing Hyperscale Campuses to Texas. 2025. https://www.esig.energy/wp-content/uploads/2025/05/ESIG_LLTF_PresentationLancium.pdf
- 47.
Pinyon Environmental Inc. Air Quality Permit by Rule. Texas Commission on Environmental Quality (TCEQ). 2025.
- 48.
Spenrath M, Datta A. Powering Texas’ digital economy: Data centers and the future of the grid: Part two. The Woodlands, Texas: Houston Advanced Research Center (HARC). 2025. https://harcresearch.org/research/powering-texas-digital-economy-data-centers-and-the-future-of-the-grid/