SAY NO TO DATA CENTERS · RESEARCH FILE
The evidence behind the initiative
Thirty evidence cards. Eleven source publications. A case for stopping new data center construction until communities have credible water and power plans, enforceable cost protections, and a meaningful say.
Selected and organized by Emily Liu. This collection brings together academic studies, energy analysis, and official public records. Each card pairs a published finding with the initiative’s argument for a pause.
Water protections come first · The infrastructure behind every facility · Grid readiness and who pays · Public costs and neighborhood health · Community control before construction
The pause recommendations are this initiative’s conclusions. A citation does not imply that its authors endorse a moratorium. Some publications support several distinct cards; projections and regional findings retain their stated scope.
RESEARCH SECTION 1 / 5
Water protections come first
01 / California policy research
Protect water before breaking ground
The report identifies site selection and facility design as the largest opportunities for water savings.
Our case for a pause: Stop construction until water-saving requirements shape the design.
Grimm, Green Nylen & Kiparsky · 2026
Credibility: Berkeley’s Wheeler Water Institute: Grimm is a research fellow, Green Nylen a senior fellow, and Kiparsky its director.
02 / California policy research
Location is a water decision
Climate, water availability, and competing uses influence a facility’s water impacts.
Our case for a pause: Reject unsuitable sites before a permanent facility commits local resources.
Grimm, Green Nylen & Kiparsky · 2026
Credibility: Berkeley’s Wheeler Water Institute: Grimm is a research fellow, Green Nylen a senior fellow, and Kiparsky its director.
03 / California policy research
No approval without water disclosure
Missing facility-level information prevents reliable assessment of local water impacts.
Our case for a pause: Require public water estimates before construction approval.
Grimm, Green Nylen & Kiparsky · 2026
Credibility: Berkeley’s Wheeler Water Institute: Grimm is a research fellow, Green Nylen a senior fellow, and Kiparsky its director.
04 / Texas research summary
Texas planning must catch up
Cook explains that historical-use planning can miss new data center demand until facilities are operating.
Our case for a pause: Pause new construction until regional plans account for proposed demand.
Margaret Cook, PhD · HARC · 2026
Credibility: Cook leads Water and Community Resilience at the Houston Advanced Research Center.
05 / Texas research summary
Measure demand before promising supply
HARC recommends reporting both expected and actual electricity and water use.
Our case for a pause: Make independently reviewable demand estimates a condition of moving forward.
Margaret Cook, PhD · HARC · 2026
Credibility: Cook leads Water and Community Resilience at the Houston Advanced Research Center.
06 / Texas research summary
Require less thirsty designs
HARC identifies reuse, alternative water supplies, and more efficient cooling as options to reduce pressure.
Our case for a pause: Require an alternatives assessment before committing drinking-water supplies.
Margaret Cook, PhD · HARC · 2026
Credibility: Cook leads Water and Community Resilience at the Houston Advanced Research Center.
RESEARCH SECTION 2 / 5
The infrastructure behind every facility
07 / Modeling study · preprint
Plan for the hottest day
The study finds peak cooling withdrawals can exceed what host water systems can supply.
Our case for a pause: Require peak-day capacity studies before construction.
Han, Li, Wierman & Ren · 2026
Credibility: Coauthors include Caltech professor Adam Wierman and UC Riverside engineering professor Shaolei Ren.
08 / Modeling study · preprint
Efficiency promises do not erase demand
Even with annual 10% water-intensity improvements, modeled new U.S. capacity needs reach 227–604 million gallons daily by 2030.
Our case for a pause: Pause projects lacking demonstrated water capacity.
Han, Li, Wierman & Ren · 2026
Credibility: Coauthors include Caltech professor Adam Wierman and UC Riverside engineering professor Shaolei Ren.
09 / Modeling study · preprint
Water savings can shift stress to power
The authors warn that dry cooling can increase electricity pressure during summer peaks.
Our case for a pause: Evaluate water and power together before approving a design.
Han, Li, Wierman & Ren · 2026
Credibility: Coauthors include Caltech professor Adam Wierman and UC Riverside engineering professor Shaolei Ren.
10 / Peer-reviewed historical modeling
Count the water used to make electricity
The study accounts for cooling water and water used at power plants; nearly half of servers received some power from water-stressed regions.
Our case for a pause: Require a complete water footprint before construction.
Siddik, Shehabi & Marston · 2021
Credibility: Peer-reviewed Environmental Research Letters study; coauthor Arman Shehabi is a Berkeley Lab staff scientist specializing in energy and environmental impacts.
11 / Peer-reviewed historical modeling
Keep new demand away from stressed watersheds
The historical model attributes one-fifth of servers’ direct water footprint to moderately or highly stressed watersheds.
Our case for a pause: Screen proposed sites for existing scarcity and competing needs.
Siddik, Shehabi & Marston · 2021
Credibility: Peer-reviewed Environmental Research Letters study; coauthor Arman Shehabi is a Berkeley Lab staff scientist specializing in energy and environmental impacts.
12 / Global energy outlook
Demand is growing faster than reassurance
The IEA projects global data center electricity use rising from 485 TWh in 2025 to 950 TWh in 2030.
Our case for a pause: Require a credible local supply plan before construction.
International Energy Agency · 2026
Credibility: The IEA is an intergovernmental agency specializing in energy statistics and scenario analysis.
RESEARCH SECTION 3 / 5
Grid readiness and who pays
13 / Global energy outlook
More efficient tasks can still mean more power
Energy per task is falling, while usage and energy-intensive applications are expanding.
Our case for a pause: Judge projects by total demand, not efficiency claims alone.
International Energy Agency · 2026
Credibility: The IEA is an intergovernmental agency specializing in energy statistics and scenario analysis.
14 / Global energy outlook
Infrastructure cannot be assumed ready
The IEA reports tightening supply-chain bottlenecks and planning systems stretched by project applications.
Our case for a pause: Stop construction where essential infrastructure readiness remains unproven.
International Energy Agency · 2026
Credibility: The IEA is an intergovernmental agency specializing in energy statistics and scenario analysis.
15 / Comparative grid-policy research
Transmission takes longer than buildings
The report contrasts 18–24-month data center construction with 7–10-year high-voltage transmission development.
Our case for a pause: Align construction permission with realistic grid delivery.
Mural, Pherwani & Gupta · 2026
Credibility: Their analysis is published by Harvard’s Belfer Center and the Harvard Paulson School’s Power and AI Initiative.
16 / Comparative grid-policy research
Speculative projects distort planning
Duplicate service requests can inflate demand forecasts and complicate infrastructure decisions.
Our case for a pause: Require firm commitments before authorizing costly buildouts.
Mural, Pherwani & Gupta · 2026
Credibility: Their analysis is published by Harvard’s Belfer Center and the Harvard Paulson School’s Power and AI Initiative.
17 / Comparative grid-policy research
On-site power still needs scrutiny
Co-location can divert supply, while grid backup arrangements raise questions about who funds standby capacity.
Our case for a pause: Resolve community reliability and backup costs before construction.
Mural, Pherwani & Gupta · 2026
Credibility: Their analysis is published by Harvard’s Belfer Center and the Harvard Paulson School’s Power and AI Initiative.
18 / Utility announcement
Households need binding cost protection
AEP reports approved requirements for large new data centers to pay for at least 85% of subscribed energy, even when using less.
Our case for a pause: Require enforceable cost safeguards first.
AEP Ohio · July 2025
Credibility: AEP Ohio operates the affected utility; this is its firsthand account of an approved tariff, with a commercial perspective.
RESEARCH SECTION 4 / 5
Public costs and neighborhood health
19 / Utility announcement
Protect residents if a project disappears
The approved arrangement includes financial-viability requirements and exit fees for cancellation or unmet obligations.
Our case for a pause: Require financial security before infrastructure is built.
AEP Ohio · July 2025
Credibility: AEP Ohio operates the affected utility; this is its firsthand account of an approved tariff, with a commercial perspective.
20 / Utility announcement
A utility pause can buy time
AEP had paused new Central Ohio data center agreements while regulators considered protective rules.
Our case for a pause: Use a targeted pause until enforceable conditions exist.
AEP Ohio · July 2025
Credibility: AEP Ohio operates the affected utility; this is its firsthand account of an approved tariff, with a commercial perspective.
21 / University summary of original research
Backup generators have a real footprint
The study found Northern Virginia data center particulate emissions rose 139% between 2015 and 2023.
Our case for a pause: Require cumulative emissions review before adding facilities.
Pitt, Suen & Plisko · VCU · 2026
Credibility: Damian Pitt is a VCU public-affairs associate professor; the team analyzed state emissions records for 138 data centers.
22 / University summary of original research
Permitted emissions leave room for escalation
Actual emissions were only about 4–7% of permitted totals. Permit ceilings describe potential exposure, not current pollution.
Our case for a pause: Evaluate maximum permitted operations before approving construction.
Pitt, Suen & Plisko · VCU · 2026
Credibility: Damian Pitt is a VCU public-affairs associate professor; the team analyzed state emissions records for 138 data centers.
23 / University summary of original research
Pollution is not evenly distributed
Higher carbon-monoxide and nitrogen-oxide emissions correlated with lower-income and more renter-heavy areas in this regional study.
Our case for a pause: Require neighborhood-level impact review before construction.
Pitt, Suen & Plisko · VCU · 2026
Credibility: Damian Pitt is a VCU public-affairs associate professor; the team analyzed state emissions records for 138 data centers.
24 / Research-institute policy explainer
Count long-term public obligations
WRI warns that infrastructure can outlast a facility, leaving communities exposed to unrecovered costs.
Our case for a pause: Require funded closure and cost-recovery commitments before building.
Carla D. Walker & Ian Goldsmith · 2026
Credibility: Walker directs U.S. Environmental Justice and Equity at WRI; Goldsmith is its clean-energy research associate.
RESEARCH SECTION 5 / 5
Community control before construction
25 / Research-institute policy explainer
Protect public access to project details
WRI identifies nondisclosure agreements as a barrier to understanding proposed projects and their resource demands.
Our case for a pause: Publish essential project information before approval.
Carla D. Walker & Ian Goldsmith · 2026
Credibility: Walker directs U.S. Environmental Justice and Equity at WRI; Goldsmith is its clean-energy research associate.
26 / Research-institute policy explainer
Land has competing community uses
The analysis describes pressure from large campuses on agricultural land and other local priorities.
Our case for a pause: Require transparent comparisons of sites and community alternatives.
Carla D. Walker & Ian Goldsmith · 2026
Credibility: Walker directs U.S. Environmental Justice and Equity at WRI; Goldsmith is its clean-energy research associate.
27 / Official local ordinance
Set noise limits before construction
Fairfax’s 2024 amendment requires a noise study before site-plan approval and another after construction.
Our case for a pause: Require enforceable acoustic protections before work begins.
Fairfax County Board of Supervisors · 2024
Credibility: The county’s adopted ordinance is a primary record of its own land-use requirements.
28 / Official local ordinance
Give homes meaningful separation
The amendment specifies a 200-foot residential setback for buildings, with exceptions requiring special approval.
Our case for a pause: Establish protective siting rules before new campuses are built.
Fairfax County Board of Supervisors · 2024
Credibility: The county’s adopted ordinance is a primary record of its own land-use requirements.
29 / Official county record
A pause is a practical policy tool
In July 2026, Linn County approved an 18-month pause on new large-scale data center rezoning applications in unincorporated areas.
Our case for a pause: Consider a defined local pause with clear review goals.
Linn County Board of Supervisors · 2026
Credibility: The county’s official record establishes the scope and purpose of its own zoning decision.
30 / Official county record
Study the combined burden
Linn County’s pause provides time for technical studies, utility coordination, public input, and cumulative-impact review.
Our case for a pause: Complete regional assessment before committing to further construction.
Linn County Board of Supervisors · 2026
Credibility: The county’s official record establishes the scope and purpose of its own zoning decision.
The source shelf
Open the original publications to examine methods, assumptions, and context. The 2021 footprint study is historical; the water-capacity paper is a preprint; local policies demonstrate available tools within their own jurisdictions.
Regulating Data Center Water Use in California — Grimm, Green Nylen & Kiparsky · 2026
Thirsty Data: Texas’ Water and Energy Costs — Margaret Cook, PhD · HARC · 2026
Small Bottle, Big Pipe — Han, Li, Wierman & Ren · 2026
The Environmental Footprint of U.S. Data Centers — Siddik, Shehabi & Marston · 2021
Key Questions on Energy and AI — International Energy Agency · 2026
Virginia and Texas: Large-Scale Electric Growth — Mural, Pherwani & Gupta · 2026
Data Center Cost Protections Approved by PUCO — AEP Ohio · July 2025
Northern Virginia Generator Pollution Study — Pitt, Suen & Plisko · VCU · 2026
How Data Centers Affect U.S. Communities — Carla D. Walker & Ian Goldsmith · 2026
Adopted Data Center Zoning Amendment — Fairfax County Board of Supervisors · 2024
Official Temporary Moratorium Announcement — Linn County Board of Supervisors · 2026
Our ask: pause construction. Put protections first.
We support stopping new data center construction until independent review shows that local water supplies, grid reliability, public finances, and neighboring communities are protected. Publish the studies, give residents time to respond, and make the conditions enforceable before work begins.
