American data centers in just seven states are reportedly using more water than all the cities in California

Summary

Data centers across seven US states rely on power generation that withdraws 3.4 trillion gallons of freshwater annually, according to Ceres. However, industry analysts note that gross withdrawal figures include water returned directly to waterways, meaning net water consumption is substantially lower.

Key Takeaways

1. Ceres estimates data centers use 3.4 trillion gallons of freshwater a year, or 10.4 million acre-feet.
2. It measures water withdrawals, or the total volume of water a power company pulls from a water source like a river, lake, or aquifer, including water it essentially replaces shortly afterward.
3. At a time when the two largest Colorado River reservoirs have hit record lows, Ceres noted that local opposition grounded in water and grid concerns disrupted $130 billion of data center projects in the first quarter of 2026 alone.

Statement Breakdown

• Claimed Facts: 65% of statements the article presents as facts
• Opinions: 25% of statements classified as editorial or subjective
• Claims: 10% of statements surfaced for additional reader evaluation

Credibility & Bias Reasoning

Credibility assessment: The report cites verifiable figures from sustainability think tank Ceres, the California Legislative Analyst's Office, and academic research. It provides critical nuance by distinguishing between total water withdrawal and actual net consumption. The reporting contextualizes high-level estimates alongside contrasting academic benchmarks.

Bias assessment: Technical Policy Analysis. The piece maintains an analytical and clarifying perspective focused on methodology distinctions. It evaluates potential overcounting in think tank findings while acknowledging valid resource constraints. The tone remains informative and grounded in infrastructure analysis.

Note: Note the technical distinction highlighted between gross water withdrawal and net consumption when evaluating data center impacts.

Credibility flag: Methodological Context

Claimed Facts (5)

• States a factual event regarding the publication and geographical scope of the research.
• Cites a documented comparison and public agency dataset.
• Defines the specific accounting methodology of the California municipal baseline.
• Provides baseline academic research findings on nationwide data center consumption.
• Documents specific capital delays linked to resource constraints in early 2026.

Opinions (5)

• Offers an analytical interpretation comparing competing dataset magnitudes.
• Reflects an editorial evaluation of the research organization's methodology.
• Expresses an interpretive judgment regarding the characterization of hydroelectric resource use.
• Relays a stakeholder perspective advocating for collaborative problem solving.
• Offers a speculative concluding observation on industry investment and artificial intelligence expansion.

Claims (5)

• Questionable whether the top-line framing fairly represents consumption without immediate context.
• Uses withdrawal numbers in a manner that can easily be misread as net resource depletion.
• Broad claim grouping diverse utility footprints into sweeping vulnerability categories.
• Forward-looking modeling based on aggressive computational demand projections.
• Asserts operational gross dependence directly equates to local community risk without utility-specific routing.

Key Sources

• Ceres — Sustainability non-profit research think tank
• Rahim Amir — Technology and infrastructure writer at TechRadar Pro
• Bobby Magill — Journalist at Bloomberg Law
• Legislative Analyst's Office — California state nonpartisan policy and fiscal analysis agency
• Arizona State University — Academic research institution conducting power and water modeling
• Shama Perveen — Director of Water Research at Ceres and report co-author