2026-06-25

A Looming Water Crisis:

The Prisoner’s Dilemma

A Policy & Legal Analysis—June 2026

Introduction

The United States is confronting a water crisis of historic proportions—and the legal, political, and institutional architecture designed to address such crises is failing on multiple fronts. This is not merely a story about drought, overuse, or climate change. It is a story about structural democratic failure: how gerrymandered legislatures, a narrowing Supreme Court, dismantled regulatory authority, and the cold logic of game theory combine to leave one of humanity’s most essential resources without an adequate guardian.

American water law is a fractured inheritance. Eastern states operate under riparian rights—rooted in English common law—in which landowners adjacent to water share reasonable use of the water as a community resource. Western states operate under prior appropriation: “first in time, first in right,” a doctrine born in the gold rush that treats water as a property right governed by the command “use it or lose it.” These two systems, designed for different geographies and different eras, collide uneasily with the demands of a 21st-century economy under climatic stress.

Into this already strained system has come a new and voracious consumer: the data center. U.S. data centers directly consumed 17.4 billion gallons of water in 2023—equivalent to the daily needs of roughly 160,000 American households—and that figure is projected to reach between 38 and 73 billion gallons by 2028¹. Driven by the explosive growth of artificial intelligence, these facilities are being built at speed and scale, overwhelmingly in water-stressed regions.

“Every thread of this crisis leads back to the same structural knot: a democracy whose inputs have been corrupted cannot produce the outputs a crisis demands.”

I. The Regulatory Vacuum

Federal authority over water rests on the Commerce Clause and on landmark statutes: the Clean Water Act, the Safe Drinking Water Act, and a web of agency rules built atop these foundations. That architecture is now being systematically dismantled by the Supreme Court at precisely the moment the crisis demands it most.

In West Virginia v. EPA (2022), the Court’s 6-3 majority explicitly adopted the “Major Questions Doctrine,” holding that agencies cannot claim authority over matters of “vast economic and political significance” without clear and explicit congressional authorization². The ruling neutered the EPA’s ability to impose broad greenhouse gas regulations—and its logic extends to virtually every major environmental regulatory action.

One year later, in Sackett v. EPA (2023), the Court applied similar textualist logic to the Clean Water Act, ruling that federal jurisdiction over wetlands requires a continuous surface connection to navigable waters³. The ruling removed federal protection from roughly half of American wetlands overnight. In Loper Bright Enterprises v. Raimondo (2024), the Court overruled the forty-year-old Chevron doctrine entirely, stripping agencies of deference on their own statutory interpretations⁴.

The cumulative effect is a three-way institutional failure. Courts say gerrymandering is a political question—go to Congress. Congress is gridlocked. Agencies attempt to fill gaps through regulation—courts say not without explicit congressional authorization. The escape valves are being closed one by one.

II. The Gerrymandering Connection

The word “gerrymandering” does not appear in the Constitution. No amendment authorizes it. No federal statute created it. It exists in a legal vacuum —neither authorized nor prohibited by any enforceable federal mechanism. The Supreme Court codified that vacuum in Rucho v. Common Cause (2019), when Chief Justice Roberts held for a 5-4 majority that partisan gerrymandering claims present “political questions beyond the reach of the federal courts”⁵.

The consequence is a nearly closed loop of unaccountability. Legislators draw maps to protect their own power. Those maps produce legislatures that serve donor bases rather than median voters. Those legislatures cannot pass reform. Courts will not intervene. And the laws that emerge from this corrupted democratic process—on water, on the environment, on anything—reflect concentrated interests, not the public good.

The connection to water is direct. Western congressional delegations from safely gerrymandered districts serve powerful agricultural water users who are among their largest donors. Legislators have no electoral incentive to cut allocations that donors depend on. Reform proposals die in committee. Consider the Colorado River Compact of 1922—a governing agreement negotiated during an unusually wet period, before climate science, before modern hydrology, before the Sun Belt. It allocates more water than the river actually produces in most years. Seven states and Mexico depend on it. And the Law of the River—the constellation of compacts, decrees, and statutes built upon it—is being applied to a 21st-century crisis it was never designed to address.

“The Colorado River Compact allocates more water than the river produces. The legal system’s answer has been to remove the tools that might fix it.”

III. The Prisoner’s Dilemma

Game theory offers the clearest lens for understanding why state-by-state water reform—the only politically viable route after Rucho closed the federal courthouse door—is so difficult to achieve.

In the classic Prisoner’s Dilemma, two rational actors both defect even though mutual cooperation would produce the best collective outcome. If one cooperates and the other defects, the cooperator suffers the worst possible outcome. Rational self-interest drives both to defect.

Applied to water: if all states agreed to fair allocation standards and independent regulatory frameworks, all would benefit from sustainable long-term water security. But if one state imposes strict water regulations while neighboring states do not, industry—including data centers—migrates to the least-regulated jurisdiction. The reforming state loses economic activity without gaining regional water security.

This is not hypothetical. Arizona has made new groundwater certificates nearly impossible for hyperscale data center projects. Virginia has imposed mandatory reclaimed water use clauses. But Texas has imposed no such constraints, and a University of Houston study projects that Texas data centers will consume 49 billion gallons of water in 2025, rising to as much as 399 billion gallons by 2030—equivalent to draining Lake Mead by more than 16 feet annually⁶.

The prisoner’s dilemma has known solutions: a binding simultaneous agreement, repeated-game punishment mechanisms, a third-party enforcer, or changed payoff structures. Every solution runs into the same wall. A binding federal agreement requires a functional Congress. Repeated-game punishment operates on 10-year redistricting cycles—too slow to deter. The federal court enforcer withdrew in Rucho and Sackett. And changing payoffs requires political will that safe districts insulate legislators from having to generate.

IV. The Data Center Accelerant

Data centers have transformed from a background infrastructure concern into an acute water crisis accelerant. A single large hyperscale facility can consume between one and five million gallons of water per day—equivalent to the daily needs of a city of 50,000 people⁷. AI workloads have increased cooling demand three to six times since 2022. Two-thirds of new hyperscale campuses built since 2022 are located in high or extreme water-stress counties⁸.

The pollution dimension compounds the consumption problem. Cooling water returned to water systems carries higher concentrations of dissolved solids—calcium, chloride, silica—that affect drinking water quality, depress crop yields, and are toxic to aquatic ecosystems. Improperly maintained cooling towers can harbor Legionella bacteria. Fewer than one-third of data center operators track water consumption, and some local government officials have signed non-disclosure agreements with developers, shielding this information from the public⁹.

Technical solutions exist and are advancing. Microsoft’s closed-loop cooling systems, adopted beginning in 2024, can save more than 33 million gallons of coolant per facility per year by circulating it without evaporative loss. Water reuse systems drawing on treated wastewater and greywater can reduce freshwater consumption by up to 80%¹⁰. But technology without mandate is voluntary. And voluntary in a prisoner’s dilemma means that whoever imposes costs on themselves first loses ground to those who do not.

V. What Reform Requires

A genuine remedy to the water crisis requires simultaneous action on three planes: democratic, legal, and technical. The interconnection is not incidental—it is structural.

On the democratic plane, gerrymandering reform is the load-bearing prerequisite. A federal mandate for independent redistricting commissions—authorized by the Elections Clause of Article I, Section 4—would not require creating a new constitutional right. It would merely fill a prohibition gap that Rucho explicitly said Congress could fill. Conditional preemption—the mechanism by which federal law sets a floor while states choose how to comply or face federal fallback commissions—is constitutionally established and well-tested in environmental law.

On the legal plane, a comprehensive federal water framework law must be drafted with sufficient specificity to withstand Major Questions scrutiny. The lesson of West Virginia v. EPA is not that the EPA cannot regulate—it is that Congress must say so clearly. A statute explicitly authorizing federal minimum water standards for data centers, establishing interstate water compacts with enforcement mechanisms, and restoring wetland protections stripped by Sackett would be legally durable if drafted with sufficient precision.

On the technical plane, mandatory Water Usage Effectiveness standards, phased timelines for transitioning to closed-loop cooling, water offset requirements for facilities in stressed watersheds, and mandatory public disclosure of consumption data would transform voluntary best practices into enforceable minimums. None requires constitutional amendments. All require a Congress capable of legislating.

“The tools to fix this crisis exist—constitutionally, statutorily, and technologically. What is missing is the democratic infrastructure to deploy them.”

Conclusion

The water crisis and the democracy crisis are not separate problems. They are the same problem. Gerrymandering produces legislators who serve concentrated interests over the public good. A Court narrowing agency authority closes the regulatory escape valves. Game theory ensures states cannot cooperate unilaterally to fill the void. And into that void, data centers—the physical infrastructure of the AI economy—pour billions of gallons of irreplaceable water each year.

Common law and purposivist judicial interpretation can right individual wrongs and remedy discrete harms. But systemic crises of this magnitude—a depleting Colorado River, a groundwater table in free fall, an AI economy with the water footprint of a mid-sized nation—require legislative action at scale. That action requires a functional Congress. A functional Congress requires fair elections. Fair elections require ending gerrymandering.

The prisoner’s dilemma has one reliable solution: a trusted enforcer who binds all players simultaneously, preventing any unilateral defection. In the American constitutional system, that enforcer is Congress. When Congress cannot act, the dilemma has no solution within the existing architecture. That is where the United States finds itself today—with the water running out and the instruments of remedy locked inside the very problem they are meant to solve.

Footnotes

Sources

Case Law (via official court records and SCOTUSblog)

Rucho v. Common Cause, 588 U.S. 684 (2019)—supremecourt.gov | brennancenter.org | scotusblog.com

West Virginia v. EPA, 597 U.S. 697 (2022)—supremecourt.gov | en.wikipedia.org/wiki/West_Virginia_v._EPA

Sackett v. EPA, 598 U.S. 651 (2023)—supreme.justia.com/cases/federal/us/598/21-454/

Loper Bright Enterprises v. Raimondo, 603 U.S. ___ (2024)—overruling Chevron U.S.A. v. NRDC, 467 U.S. 837 (1984)

Marbury v. Madison, 5 U.S. 137 (1803)—foundational judicial review precedent

Printz v. United States, 521 U.S. 898 (1997)—anti-commandeering doctrine

Smiley v. Holm, 285 U.S. 355 (1932)—Congress authority to override state election rules

Government Reports & Data

Shehabi, A. et al. (2024). 2024 United States Data Center Energy Usage Report. Lawrence Berkeley National Laboratory. LBNL-200163. doi.org/10.71468/P1WC7Q

U.S. Congressional Research Service (2026). Data Centers and Their Energy Consumption: Frequently Asked Questions. R48646. congress.gov/crs-product/R48646

U.S. EPA (2025). Data center water consumption statistics cited in MOST Policy Initiative (April 2026).

Academic Sources

Bondurant, E.J. (2021). “Rucho v. Common Cause—A Critique.” Emory Law Journal 70(5). scholarlycommons.law.emory.edu/elj/vol70/iss5/1/

Joyce, S.B. (2024). “Testing the Major Questions Doctrine.” Stanford Environmental Law Journal 43:50. law.stanford.edu/publications/testing-the-major-questions-doctrine/

Privette, C. et al. (2026). “Data Centers Water Footprint: The Need for More Transparency.” AGU Advances. agupubs.onlinelibrary.wiley.com/doi/10.1029/2025AV002140

Harvard Law Review (2019). “Political Question Doctrine—Rucho v. Common Cause.” harvardlawreview.org/wp-content/uploads/2019/11/252-261_Online.pdf

Georgetown Environmental Law Review (2023). “Loaded Docket: Sweeping Constitutional Challenges Facing Regulatory Authority.” law.georgetown.edu/environmental-law-review/blog/loaded-docket

American Bar Association (2023). “The Major Questions Doctrine Post-West Virginia v. EPA.” americanbar.org/groups/environment_energy_resources/resources/trends/2023-january-february/

**Industry & Policy Reports **

MOST Policy Initiative. “Data Center Water Use.” April 8, 2026. mostpolicyinitiative.org/science-note/data-center-water-use/

Lincoln Institute of Land Policy. “Data Drain: The Land and Water Impacts of the AI Boom.” February 23, 2026. lincolninst.edu/publications/land-lines-magazine/articles/land-water-impacts-data-centers/

Nature Forward. “Data Centers and Water Use.” March 20, 2026. natureforward.org/data-centers-and-water-use/

Control Associates. “Water Use and Regulation in Data Centers.” November 22, 2025. controlassociatesinc.com/blog/data-center-water-regulation-2026

AIRSYS North America. “Water Free Cooling Systems for Data Centers.” February 2026. abccarolinas.org/how-water-free-cooling-systems-are-transforming-sustainable-construction-and-data-center-design/

Common Cause. “Common Cause v. Rucho.” January 23, 2026. commoncause.org/work/common-cause-v-rucho/

Kirkland & Ellis LLP. “The Major Questions Doctrine Reigns Supreme in West Virginia v. EPA.” July 6, 2022. kirkland.com/publications/kirkland-alert/2022/07/supreme-court-west-virginia-v-epa