In 2024, Google's data center campus in Council Bluffs, Iowa consumed approximately one billion gallons of water. According to Google's Environmental Report, all of it was potable (drinking-quality water drawn from municipal supply). Google is Council Bluffs Water Works' largest customer; the company funded expansion of the city's water treatment infrastructure to support its operations.
In Storey County, Nevada, another Google data center consumed 1.5 million gallons that same year. The difference, a factor of nearly 700, comes down to a single design choice: cooling technology. Council Bluffs uses evaporative cooling. Storey County uses air cooling.
This contrast illustrates the central tension facing data center operators today. Evaporative cooling is energy-efficient but water-intensive. Air cooling eliminates most water consumption but increases electricity demand. The tradeoff is real, and it is being made, explicitly or by default, at every facility.
According to Lawrence Berkeley National Laboratory, U.S. data centers collectively consumed approximately 17 billion gallons of water for cooling in 2023. That figure could double or quadruple by 2028 as artificial intelligence workloads intensify computational demands. These numbers signal more than operational scale. They reveal a structural misalignment between the infrastructure powering the digital economy and the freshwater resources sustaining the communities where that infrastructure operates.
To understand why water has become a flashpoint, consider what a hyperscale data center represents. Industry analysts define hyperscale facilities as those containing at least 5,000 servers, occupying 10,000 square feet or more, and drawing 100 megawatts (MW) or greater of electrical power. For context, 100 MW can power approximately 80,000 U.S. households. The largest hyperscale campuses now exceed 1,000 MW, equivalent to the electrical load of 800,000 homes.
Water consumption follows power consumption, but the relationship depends on cooling system design. The dominant approach remains evaporative cooling. In these systems, water absorbs heat from servers, circulates to cooling towers, and is partially lost to the atmosphere through evaporation. A hyperscale facility using evaporative cooling typically consumes 400,000 to 550,000 gallons daily (roughly equivalent to the daily water use of 3,000 to 4,000 households).
This is consumption in the technical sense: water permanently removed from the local hydrological cycle. Approximately 80 percent evaporates into the atmosphere and cannot be recovered. The remaining 20 percent, now concentrated with dissolved minerals, must be discharged as blowdown (the periodic purge required to prevent scale formation and corrosion in cooling equipment). Both the evaporative loss and the blowdown require continuous makeup water, creating demand that can strain municipal supplies, particularly during summer months when cooling loads and community water needs spike simultaneously.
Not all data centers consume water at these rates. Closed-loop cooling systems reject heat through air-cooled chillers or dry coolers rather than evaporation. A 100 MW closed-loop facility may consume only 8,000 gallons daily, essentially limited to domestic and sanitary uses rather than cooling operations.
The tradeoff is energy. Evaporative cooling exploits a thermodynamic advantage: the phase change from liquid to vapor absorbs significant thermal energy with relatively modest electricity input. Closed-loop systems eliminate water consumption but require more electricity to achieve equivalent cooling, shifting the environmental burden from water to carbon (depending on grid mix) and increasing operating costs. Industry estimates suggest evaporative cooling reduces energy consumption by roughly 10 percent compared to air-cooled alternatives, a meaningful margin at hyperscale.
This tradeoff explains why evaporative cooling remains prevalent despite water scarcity concerns. The choice of cooling architecture is typically made during facility design and is difficult to reverse. Operators who selected evaporative systems a decade ago now face a different water landscape than the one they planned for.
Data centers cluster where three factors converge: reliable power, robust fiber connectivity, and favorable tax treatment. Water availability has historically ranked lower on site selection criteria, a calculus that has produced a geographic irony.
Across the American West, facilities have proliferated in regions already facing acute water stress. Arizona, despite limiting home construction in the Phoenix area to preserve groundwater, continues to attract data center development. Oregon's Columbia River basin hosts multiple hyperscale campuses drawing from watersheds increasingly strained by drought and competing agricultural demands. In The Dalles, Oregon – a city of 16,000 – Google's facilities accounted for more than 25 percent of the city's total water consumption in 2021.
This pattern is not accidental. The same arid climates that create water scarcity also offer lower humidity, which improves cooling efficiency and enables more hours of "free cooling" using outside air. Cheap land, tax incentives, and proximity to renewable energy sources further concentrate development in water-stressed corridors. The result: industrial water users and municipalities increasingly compete for the same constrained resource, often without coordinated planning.
The consequences of this geographic mismatch are no longer abstract. According to Data Center Watch, proposed projects representing approximately $64 billion in investment have been delayed, blocked, or abandoned due to community opposition and regulatory pushback with water concerns frequently cited alongside energy and noise objections.
Local resistance has intensified as communities recognize the scale of resource competition. In regions where aquifers are declining and agricultural users face curtailment, a single hyperscale facility consuming 400,000 gallons daily becomes a tangible threat to economic livelihoods and long-term water security. Municipal leaders who once welcomed data centers for their tax revenue now face constituents demanding answers about water allocation, infrastructure strain, and the distribution of benefits versus burdens.
Regulatory frameworks are evolving in response. California's legislature has considered measures requiring data center developers to provide water consumption estimates as part of permitting applications. Other jurisdictions are exploring tiered water pricing, mandatory efficiency standards, and requirements for alternative water sourcing. For operators, this regulatory momentum signals that water strategy can no longer be treated as an operational detail; moreover, it is becoming a permitting prerequisite.
Beyond local politics, a parallel force is reshaping expectations around water transparency. The CDP Water Security questionnaire has become the dominant framework for corporate water disclosure, with 4,815 companies responding in 2023, a 23 percent increase from the prior year.
For data center operators and the enterprises they serve, CDP creates a visibility chain that extends water accountability beyond facility boundaries. Companies disclosing through CDP must report not only their direct consumption but increasingly their supply chain water risks, which means hyperscale operators face pressure from enterprise customers seeking to manage their own disclosure obligations.
The bar for credibility continues to rise. Achieving CDP's A-List designation requires demonstrated progress on quantitative targets, site-level water monitoring across more than 75 percent of operations, and evidence of engagement with local watershed stakeholders. These are auditable criteria that influence investor confidence, customer procurement decisions, and competitive positioning.
The hyperscale operators have responded with ambitious public commitments. Microsoft pledged in September 2020 to become "water positive" by 2030, to replenishing more water than it consumes globally. Google made a similar commitment in September 2021; by 2024, the company reported replenishing 64 percent of its freshwater consumption through watershed stewardship projects. Amazon Web Services announced its water-positive target in November 2022.
Behind these pledges lies operational progress. Microsoft reports an 80 percent cumulative improvement in water use efficiency across successive generations of data center design, achieving water usage effectiveness (WUE) of 0.30 liters per kilowatt-hour in its newest facilities. In August 2024, Microsoft announced a next-generation architecture featuring closed-loop, chip-level cooling that virtually eliminates evaporative water loss, with pilot deployments expected in 2026.
Google has expanded its use of recycled and non-potable water sources, now employing reclaimed water at facilities in Singapore, Georgia, and elsewhere, which roughly equates to 22 percent of its data center cooling volume in 2023. Whether these innovations deliver projected results at scale remains to be seen, but the direction of investment signals recognition that the status quo is untenable.
The challenge is that water-positive commitments are inherently global, while water impacts are stubbornly local. Replenishing an aquifer in one region does not restore the watershed stressed by a facility in another. The frontier for credibility lies in demonstrating that efficiency gains and replenishment projects deliver measurable benefits to the specific communities bearing the operational burden.
For data center operators, whether hyperscale giants or enterprise facilities, the strategic question is not whether water matters, but how to manage it with the same rigor applied to power and connectivity.
That management begins with source water. Municipal supplies, groundwater, and surface water each present distinct treatment challenges based on mineral content, biological load, and seasonal variability. A facility drawing from hard groundwater faces different scaling risks than one using treated municipal supply. A campus in the arid Southwest confronts different chemistry than one in the humid Southeast.
Understanding source water quality is not an engineering nicety. It is the foundation for every downstream decision about treatment chemistry, cycles of concentration (the ratio of dissolved minerals in circulating water compared to makeup water), and blowdown management – the operational mechanics that determine both water efficiency and discharge compliance. Part 2 of this series examines these tradeoffs in detail.
Water is no longer a background utility for data center operations. It is a strategic variable that shapes permitting timelines, community relationships, regulatory exposure, and increasingly, customer procurement decisions.
The operators who treat water management as a core competency, investing in source water expertise, optimizing cooling system efficiency, and engaging transparently with the communities they depend on, will find themselves better positioned for expansion and better insulated from regulatory friction.
Those who continue treating water as someone else's problem may find that the communities and regulators they depend on have reached a different conclusion.
The question is no longer whether data centers will face water constraints. It is whether operators will adapt before those constraints become competitive liabilities.
Next in this series: Part 2 examines the chemistry tightrope, how cycles of concentration, blowdown optimization, and cooling system choices determine water efficiency and shape discharge compliance. Part 3 explores wastewater as a strategic asset, from Clean Water Act considerations to emerging reuse pathways.
Harrison Lee is Vice President of Strategic Marketing at Valicor Environmental Services, where he focuses on the intersection of industrial water management, sustainability communications, and environmental compliance. This article is part of a series examining water challenges in data center operations.
1. Google Environmental Report (2025). Water consumption data for Council Bluffs and Storey County. https://sustainability.google/reports/google-2025-environmental-report/
2. Lawrence Berkeley National Laboratory (2024). United States Data Center Energy Usage Report. https://eta.lbl.gov/publications/united-states-data-center-energy
3. The Conversation (2025). Data centers consume massive amounts of water – companies rarely tell the public exactly how much. https://theconversation.com/data-centers-consume-massive-amounts-of-water-companies-rarely-tell-the-public-exactly-how-much-262901
4. Council Bluffs Water Works Annual Reports (2020-2023). https://www.cbwaterworks.com/meetings-and-reports/annual-report/
5. Google Water Stewardship Report (2021). https://blog.google/outreach-initiatives/sustainability/replenishing-water/
6. Microsoft Official Blog (September 2020). Microsoft will replenish more water than it consumes by 2030. https://blogs.microsoft.com/blog/2020/09/21/microsoft-will-replenish-more-water-than-it-consumes-by-2030/
7. AWS (November 2022). Water Positive Commitment announcement. https://press.aboutamazon.com/2022/11/aws-makes-water-positive-commitment-to-return-more-water-to-communities-than-it-uses-by-2030
8. CDP Water Security. https://www.cdp.net/en/water
9. HDR Engineering (2022). Council Bluffs Water Works Council Point Water Treatment Plant. https://www.hdrinc.com/portfolio/council-bluffs-water-works-council-point-water-treatment-plant
10. Google Data Centers: Locations. https://datacenters.google/locations/
11. ESG Today (September 2021). Google Commits to be Water Positive. https://www.esgtoday.com/google-commits-to-be-water-positive-replenishing-more-water-than-consumed-by-2030/
12. Data Centre Magazine (2025). Google Environmental Report: Data Centre Impact. https://datacentremagazine.com/news/google-environmental-report-2025-the-data-centre-impact