Data Center Watch (a research project from 10a Labs) · undated on page; research timeline printed as May 2024 - March 2025 · Web report
The first Data Center Watch report catalogs local opposition to data center projects in the 28 states where hyperscalers have projects of 50 MW or more in planning. Its methodology section says it built a database of activist groups and public officials from local media, government filings, petitions, social media and official statements. It defines 'blocked' and 'delayed' projects, then profiles 16 specific cases, including Peculiar, Missouri, where the Planning Commission removed data centers from the zoning ordinance in October 2024, and the PW Digital Gateway in Prince William County, Virginia. It also covers bipartisan opposition, the Virginia Data Center Reform Coalition, recall and council elections (Cascade Locks, Oregon and Warrenton, Virginia), and petition tracking.
“$18 billion worth of data center projects were blocked, and another $46 billion of projects were delayed over the last two years in the face of opposition from residents and activist groups.” (Executive Summary)
“There are at least 142 activist groups across 24 states organizing to block data center construction and expansion.” (Executive Summary)
“some common themes are higher utility bills, water consumption, noise, impact on property value, and green space preservation.” (Executive Summary)
“55% of the politicians who had taken public positions against the data center projects were Republicans, and 45% were Democrats.” (Executive Summary)
Who should read it:
Local elected officials, planning commissioners and economic development staff who want to understand why projects stall and what concerns residents raise; especially useful for Missouri officials because of the Peculiar case study.
Limitations:
The report itself says attributing delays exclusively to opposition 'oversimplifies a complex landscape' and that 'some figures reflect estimates that may change.' It states it does not assess the merits of activist concerns. Data Center Watch is run by 10a Labs; the page describes its research as 'objective, fact-based, and nonpartisan' but the report frames opposition as a development risk.
Cite as:
Data Center Watch (a research project from 10a Labs). “$64 billion of data center projects have been blocked or delayed amid local opposition.” undated on page; research timeline printed as May 2024 - March 2025. https://www.datacenterwatch.org/report
Data Center Watch (10a Labs) · undated on page; research timeline printed as Late March 2025 - June 2025 · Web summary page
This quarterly update summarizes opposition activity from late March to June 2025. The public page gives key takeaways only: the number of projects blocked or delayed, the dollar value affected, growth in the count of active opposition groups, and the rise of tax abatement rollbacks as a political risk (citing suspended projects in Minnesota and South Dakota). It notes petitions, public hearings and grassroots organizing reshaping approval processes, especially in Indiana and Georgia.
“In just three months, 20 projects were blocked or delayed amid local opposition, affecting $98 billion in potential investment” (Key Takeaways)
“Community opposition continues to grow, with 53 active groups across 17 states targeting 30 data center projects in Q2 alone, bringing the total to 188 groups nationwide.” (Key Takeaways)
“During this period, 66% of the tracked protested projects were blocked or delayed.” (Key Takeaways)
Who should read it:
Officials tracking how quickly local opposition is growing and how tax abatement debates affect project risk.
Limitations:
Only a summary is public; the page says 'Interested in purchasing the report?' Same publisher caveats as the first Data Center Watch report.
Cite as:
Data Center Watch (10a Labs). “Data Center Watch Report Q2 2025 UPDATE: 125% Surge in Data Center Opposition.” undated on page; research timeline printed as Late March 2025 - June 2025. https://www.datacenterwatch.org/q22025
Data Center Watch (10a Labs) · undated on page; research timeline printed as July 2025 - December 2025 · Web summary page
This update covers the second half of 2025 and argues that community opposition is now a core element of data center development rather than an anecdotal issue. Its key takeaways describe opposition moving into litigation, moratoria and challenges to by-right development, and into state legislatures, utility regulation and elections, naming Virginia, New Jersey and Georgia.
“In 2025, local opposition blocked or delayed dozens of data center projects representing $152B in potential investment.” (Key Takeaways)
Who should read it:
Officials weighing whether to keep data centers as a by-right use or require legislative approval.
Limitations:
Summary page only. The page title in the browser still read 'Q2 2025 UPDATE' although the body is the Q3-Q4 update. Same publisher caveats as above.
Cite as:
Data Center Watch (10a Labs). “Q3-Q4 2025 UPDATE: Data center opposition has consolidated.” undated on page; research timeline printed as July 2025 - December 2025. https://www.datacenterwatch.org/q3-q4-2025
Data Center Watch (10a Labs) · undated on page; research timeline printed as January 2026 - March 2026 · Web summary page
The most recent Data Center Watch update reports the largest single-quarter concentration of blocked and delayed projects it has recorded. It describes a 'structural shift' in which communities have internalized an opposition playbook, legislative sessions introduced regulatory uncertainty, and moratoria expanded from local exceptions to a multi-front political fight.
“at least 75 projects worth approximately $130 billion disrupted by local opposition” (TL;DR)
“the number of active opposition groups more than doubled across 49 states” (TL;DR)
Who should read it:
Anyone briefing a council or commission on the current national climate for data center approvals.
Limitations:
Summary page; underlying project list not public on this page. Publisher caveats as above.
Cite as:
Data Center Watch (10a Labs). “Q1 2026: Data Center Watch Report.” undated on page; research timeline printed as January 2026 - March 2026. https://www.datacenterwatch.org/q1-2026
Pew Research Center (John Gramlich, Brian Kennedy, Colleen McClain, Galen Stocking) · March 12, 2026 · Web article
Pew's first survey on data centers, fielded Jan. 20-26, 2026 among 8,512 adults, asks whether data centers are mostly good or bad for five areas: the environment, home energy costs, quality of life nearby, local jobs and local tax revenue. It breaks results down by party, ideology, age and how much people have heard about data centers. The article finds attitudes more negative than positive on environment, energy costs and quality of life, and more positive than negative on jobs and tax revenue, with a large share unsure.
“Far more say data centers are mostly bad than good for the environment (39% vs. 4%), home energy costs (38% vs. 6%) and the quality of life for those nearby (30% vs. 6%).” (Key findings list)
“More adults say data centers are mostly good than bad for local jobs (25% vs. 15%) and local tax revenue (23% vs. 12%).” (Key findings list)
“54% of adults under 30 say data centers have a mostly negative effect on the environment.” (Age differences section)
Who should read it:
Elected officials and engagement staff designing public hearings or outreach, who need a neutral national baseline on what residents expect good and bad from data centers.
Limitations:
National survey; it does not measure attitudes toward a specific local project. Around one in five or more said they were not sure in each area.
Cite as:
Pew Research Center (John Gramlich, Brian Kennedy, Colleen McClain, Galen Stocking). “How Americans view data centers’ impact in key areas, from the environment to jobs.” March 12, 2026. https://www.pewresearch.org/short-reads/2026/03/12/how-americans-view-data-centers-impact-in-key-areas-from-the-environment-to-jobs/
Gallup (Jeffrey M. Jones) · May 13, 2026 · Web article with linked PDF of question responses
Gallup's first question on local data center construction, from a March 2-18, 2026 survey, uses the same wording Gallup uses for local nuclear power plants, allowing a direct comparison. A follow-up open-ended question on an April Gallup Panel web survey asked why people favor or oppose a local data center. The article reports reasons for opposition (resource use, pollution including noise, quality of life, utility bills) and reasons for support (jobs, tax revenue), plus differences by party, gender and region.
“Seven in 10 Americans oppose constructing data centers for artificial intelligence in their local area, including nearly half, 48%, who are strongly opposed.” (Opening paragraph)
“53% of Americans say they oppose building a nuclear energy plant in their area, far less than the 71% opposed to data center construction.” (Third paragraph)
“two-thirds of those in favor of building data centers in their area cite the economic benefits, including 55% who mention increased job opportunities specifically.” (Opposition Tied to Environmental, Quality-of-Life Concerns section)
“Opposition is slightly lower among those living in the West (63%) and East (68%) than in the Midwest (76%) and South (75%).” (Majority of Democrats Strongly Oppose Data Centers section)
Who should read it:
Midwest officials in particular, since the regional breakdown shows the highest opposition in the Midwest and South; also useful for anyone drafting public engagement plans around the reasons residents give.
Limitations:
Question asks about 'data centers for artificial intelligence' specifically, which may produce different answers than a generic question.
Cite as:
Gallup (Jeffrey M. Jones). “Americans Oppose AI Data Centers in Their Area.” May 13, 2026. https://news.gallup.com/poll/709772/americans-oppose-data-centers-area.aspx
Annenberg Public Policy Center of the University of Pennsylvania · August 11, 2026 · Web press release with downloadable topline and methodology
APPC's Institutions of Democracy division surveyed 1,320 U.S. adult citizens from June 16 to July 19, 2026 and compared results with its survey ending in March 2026. The release finds that opposition to local data centers rose sharply while broader views of AI stayed stable, meaning the change is concentrated on the physical facility near people's homes. It reports results by party and age and gives the method: SSRS fielded the survey primarily online with a small phone sample, weighted to population benchmarks.
“Three in five Americans (61%) now somewhat or strongly oppose the construction of new data centers in their area, up from 49% in the survey ending in March.” (Key findings)
“Majorities of Democrats (69%), Republicans (54%) and independents (53%) oppose new local data centers.” (Key findings)
“The margin of error for the full sample is ±3.5 percentage points, and it is larger for subgroups.” (Methodology paragraph)
Who should read it:
Officials who want a trend line, not a single snapshot, of how local sentiment is moving in 2026.
Limitations:
Margin of error is stated as plus or minus 3.5 points; subgroup estimates are less precise.
Cite as:
Annenberg Public Policy Center of the University of Pennsylvania. “Opposition to Local Data Centers Rises Sharply, Annenberg Survey Finds.” August 11, 2026. https://www.annenbergpublicpolicycenter.org/opposition-to-local-data-centers-rises-sharply-annenberg-survey-finds/
UMass Amherst Poll, University of Massachusetts Amherst · September 14, 2026 · Web report with executive summary
This national poll of 1,000 respondents, fielded August 21-26, 2026, asks about support for an AI data center in the respondent's local community and about the President's handling of AI. The report has sections for an executive summary, press release, 'Views on AI Data Centers,' methodology and toplines. It lists the reasons opponents gave, including environmental or natural resource use, land use and local disruptions, distrust of AI and the industry, and utility bills.
“just 11% of Americans say they would support the construction of an artificial intelligence data center in their local community, while 65% oppose their construction.” (Executive Summary)
“even Republicans oppose a data center in their community by nearly 3-to-1, 52% to 18%.” (Executive Summary)
Who should read it:
Officials who want the most recent national reading (fielded late August 2026) before a hearing.
Limitations:
Stated margin of error 3.5%. Question wording specifies an 'artificial intelligence data center.'
Cite as:
UMass Amherst Poll, University of Massachusetts Amherst. “Just 1 in 9 Americans Support AI Data Center Construction in their Local Communities.” September 14, 2026. https://www.umass.edu/poll/about/reports/2026-09-national-public-opinion-poll-2
Information Technology and Innovation Foundation (ITIF) with Public First · August 26, 2026 · Web press release
An online survey of 2,007 U.S. adults conducted August 8-12, 2026 compared support and opposition for 10 types of local development, from new housing and roads to warehouses, power plants and data centers. It found data centers ranked last by a wide margin. It also asked whether people would rather leave land with no other viable use undeveloped than host a data center, and how support changes by the stated purpose of the data center (online banking or streaming versus AI training).
“Nearly half of Americans (46 percent) oppose new data center developments in their communities, and barely a quarter (26 percent) support them” (Opening paragraph)
“49 percent say they would prefer it to stay undeveloped and unused, and 33 percent would choose to have the land developed with a new data center.” (Body, after the comparison table)
Who should read it:
Planners comparing public acceptance of data centers with other land uses they regularly approve, such as warehouses and housing.
Limitations:
ITIF is a technology policy think tank that describes its mission as promoting policies that 'accelerate innovation'; the survey was conducted by Public First in partnership with ITIF.
Cite as:
Information Technology and Innovation Foundation (ITIF) with Public First. “Data Centers Are Americans’ Least Favored Form of Development, New Research From Public First and ITIF Shows.” August 26, 2026. https://itif.org/publications/2026/08/26/data-centers-are-americans-least-favored-form-of-development/
Loudoun County, Virginia (County Administrator) · undated on page (cites 2025 figures) · County web page
Loudoun County describes how local data centers use a mix of cooling approaches, how Loudoun Water built a reclaimed wastewater system in 2010 to serve early data centers, and how developers consult water and wastewater authorities before proposals. It notes that a special exception process adopted in March 2025 gives officials and the public more visibility into water impacts, and discusses wastewater from open-loop cooling.
“On an average day, data centers in Loudoun account for 9% of total potable water use.” (Water Use and Conservation)
“the system delivered more than 750 million gallons of reclaimed water to roughly 40 customers in 2025, directly offsetting potential use of potable water.” (Water Use and Conservation (attributed to Loudoun Water))
Who should read it:
Water utility managers and planners considering reclaimed water requirements or a discretionary review step for water-intensive proposals.
Limitations:
Figures are attributed to Loudoun Water; county page, not an independent study.
Cite as:
Loudoun County, Virginia (County Administrator). “Data Centers: Water Considerations.” undated on page (cites 2025 figures). https://www.loudoun.gov/6402/Water-Considerations
Environmental and Energy Study Institute (EESI), by Miguel Yañez-Barnuevo · June 25, 2025 · Web article
This EESI article explains how data centers use freshwater for cooling, why larger AI-focused facilities are increasing water consumption alongside energy use, and which cooling technologies (direct-to-chip and immersion) reduce water and energy use. It is part of EESI's series covering energy, water, noise and energy bills.
“Large data centers can consume up to 5 million gallons per day, equivalent to the water use of a town populated by 10,000 to 50,000 people.” (Highlights)
“Novel technologies like direct-to-chip cooling and immersion cooling can reduce water and energy usage by data centers.” (Highlights)
Who should read it:
Water utility boards and council members who need a plain-language primer on cooling and water.
Limitations:
Explainer from an environmental nonprofit that synthesizes other sources.
Cite as:
Environmental and Energy Study Institute (EESI), by Miguel Yañez-Barnuevo. “Data Centers and Water Consumption.” June 25, 2025. https://www.eesi.org/articles/view/data-centers-and-water-consumption
Pacific Institute · Published: September 2026 · Issue brief
The Pacific Institute's issue brief summarizes what is known about data center water use and addresses common misunderstandings. The landing page overview says data centers are a small share of national water use but can be locally significant, especially in water-scarce and rural areas, and that the lack of standardized reporting on data center number, size, location and water use makes impacts hard to assess.
“While data centers represent a relatively small proportion of global or national water use, data center water use can be locally significant, especially in water-scarce and rural areas.” (Overview)
“The lack of standardized reporting on the number, size, location, and water use of data centers makes it difficult for communities, water managers, and decision-makers to understand and adequately address potential impacts.” (Overview)
Who should read it:
Rural county officials and water managers who need a balanced framing of national versus local water significance.
Limitations:
Only the web overview was read; the PDF brief would not render as text in the browser.
Cite as:
Pacific Institute. “How Much Water Do Data Centers Use? And Why Does It Matter?.” Published: September 2026. https://pacinst.org/publication/how-much-water-do-data-centers-use-and-why-does-it-matter/
Ceres · September 23, 2025 · Report landing page with key findings
Grounded in the Phoenix, Arizona region, this Ceres report examines both direct on-site water use for cooling and indirect water use from the power generation that supplies data centers, and the cumulative effect of clustered facilities in water-stressed areas. It frames these as financial, operational, reputational, legal and regulatory risks and offers recommendations for companies, investors, water managers and policymakers.
“this report finds that indirect water use – primarily from power generation to meet their massive energy needs – has an even greater impact.” (Introduction)
“Water use associated with data center cooling operations is expected to increase by 870% as more facilities come online, from 385 million gallons a year to more than 3.7 billion gallons” (Key Findings)
“Data center growth could increase water stress in already strained basins by up to 17% annually” (Key Findings)
Who should read it:
Officials in arid and fast-growing regions, and anyone evaluating cumulative impacts of multiple projects rather than one site at a time.
Limitations:
Focused on the Phoenix area; Ceres is a sustainability nonprofit working with investors.
Cite as:
Ceres. “Drained by Data: The Cumulative Impact of Data Centers on Regional Water Stress.” September 23, 2025. https://www.ceres.org/resources/reports/drained-by-data-the-cumulative-impact-of-data-centers-on-regional-water-stress
Houston Advanced Research Center (HARC), Margaret Cook, PhD · 01.21.2026 (as printed) · White paper
HARC's white paper examines how current and projected data center growth could affect Texas water supplies and why state water planning may not capture that demand. The page summarizes recommendations including transparent reporting of data center water and electricity use, incentives for water-lean technology on site, and forward-looking forecasting in state water planning.
“existing and proposed data centers could place significant and growing pressure on the state’s already limited water supplies, pressure that is not fully reflected in current state water planning processes.” (Research summary)
“Texas Data Center Boom Could Consume Up to 161 Billion Gallons of Water Annually by 2030” (Title of the linked HARC press release)
Who should read it:
Officials in Texas and other states whose regional water plans do not yet account for data center demand.
Limitations:
Texas-specific. Only the landing page was read; the white paper PDF itself was not opened.
Cite as:
Houston Advanced Research Center (HARC), Margaret Cook, PhD. “Thirsty Data and the Lone Star State: The Impact of Data Center Growth on Texas’ Water Supply.” 01.21.2026 (as printed). https://harcresearch.org/research/thirsty-data-and-the-lone-star-state-the-impact-of-data-center-growth-on-texas-water-supply/
Houston Advanced Research Center (HARC), with Texas Living Waters · 07.14.2026 (as printed) · Policy recommendations
Policy recommendations for the 90th Texas Legislature on managing water demand from data centers and other large users. Partners listed are Texas Living Waters, National Wildlife Federation, Hill Country Alliance, Sierra Club Lone Star Chapter and Bayou City Waterkeeper. The recommendations are grouped in four areas: transparency and planning, responsible use, cost recovery from large users, and long-term protection of supplies and communities.
“The policy recommendations focus on four key areas: improving transparency and water-use planning, encouraging responsible use of limited water resources, ensuring large water users pay their fair share of infrastructure costs, and protecting water supplies and communities for the long term.” (Page summary)
Who should read it:
Utility and water district officials considering cost-recovery rules for large water users.
Limitations:
Developed with environmental advocacy partners; Texas-focused.
Cite as:
Houston Advanced Research Center (HARC), with Texas Living Waters. “Balancing the Water Needs of Data Centers and Large Water Users in Texas.” 07.14.2026 (as printed). https://harcresearch.org/research/balancing-the-water-needs-of-data-centers-and-large-water-users-in-texas-with-the-states-vision-for-long-term-water-security/
arXiv preprint (Pengfei Li, Jianyi Yang, Mohammad A. Islam, Shaolei Ren) · Submitted 6 Apr 2023; last revised 26 Mar 2025 (v5) · Academic paper
This widely cited academic paper estimates the water withdrawal and consumption footprint of AI models, including direct on-site evaporation for cooling. The abstract gives an example estimate for training one large language model and a projection of global AI water withdrawal in 2027, and argues that water footprint should be addressed alongside carbon.
“training the GPT-3 language model in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater” (Abstract)
“the global AI demand is projected to account for 4.2-6.6 billion cubic meters of water withdrawal in 2027” (Abstract)
Who should read it:
Staff who want the academic origin of many water figures quoted in news coverage.
Limitations:
Preprint; figures are modeled estimates, and the abstract notes such information 'has been kept a secret' by operators.
Cite as:
arXiv preprint (Pengfei Li, Jianyi Yang, Mohammad A. Islam, Shaolei Ren). “Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models.” Submitted 6 Apr 2023; last revised 26 Mar 2025 (v5). https://arxiv.org/abs/2304.03271
Cornell Chronicle, Cornell University (David Nutt) · November 10, 2025 · University news release on a peer-reviewed study
Cornell researchers built a state-by-state estimate of carbon and water impacts from AI data center growth to 2030, published November 10, 2025 in Nature Sustainability. The release reports projected annual carbon dioxide and water impacts and a 'roadmap' of smart siting, grid decarbonization and operational efficiency that would cut those impacts relative to worst-case scenarios. It identifies which regions offer the best combined carbon-and-water profile.
“by 2030, the current rate of AI growth would annually put 24 to 44 million metric tons of carbon dioxide into the atmosphere” (Second paragraph)
“It would also drain 731 to 1,125 million cubic meters of water per year” (Second paragraph)
Who should read it:
State and regional planners thinking about where large facilities should be sited.
Limitations:
University news summary of the study; the paper itself was not opened.
Cite as:
Cornell Chronicle, Cornell University (David Nutt). “‘Roadmap’ shows the environmental impact of AI data center boom.” November 10, 2025. https://news.cornell.edu/stories/2025/11/roadmap-shows-environmental-impact-ai-data-center-boom
Fairfax County, Virginia, Department of Planning and Development · Amendment adopted September 10, 2024; page undated · County web page with links to adopted text
Fairfax County's page documents its data center zoning amendment, which the Board of Supervisors directed staff to prepare on March 19, 2024 to strengthen data center permissions and standards. It links the adopted text, the staff report, follow-on motions, a separate 'Noise Study Requirements' document and the county's 'Data Centers Research Report and Recommendations' dated January 9, 2024.
“The Board of Supervisors adopted the Data Centers Zoning Ordinance Amendment on September 10, 2024, which became effective September 11, 2024.” (Adopted Text section)
Who should read it:
Zoning administrators looking for a model ordinance package, including a stand-alone noise study requirement.
Limitations:
Only the landing page was read; the linked adopted text, staff report and noise study PDFs were not opened in this session.
Cite as:
Fairfax County, Virginia, Department of Planning and Development. “Data Centers: Adopted Zoning Ordinance Amendment.” Amendment adopted September 10, 2024; page undated. https://www.fairfaxcounty.gov/planning-development/data-centers
Loudoun County, Virginia (County Administrator) · undated on page · County web page
Loudoun County, home to nearly 200 operational data centers, explains the noise sources at data center campuses (cooling, HVAC, emergency generators, on-site generation and storage), and why tonal and low-frequency noise is hard to regulate with standard A-weighted (dBA) limits. It describes the county's current rules: pre-construction and post-construction noise studies and a 55 dBA residential property-line limit, plus the complaint and enforcement process.
“For residential properties, the maximum allowable sound level at the property line is 55 dBA.” (Noise tab, zoning ordinance paragraph)
“The county does not currently have policy or regulatory language specific to tonal noise or low frequency noise.” (Noise Complaints paragraph)
Who should read it:
Planning and zoning staff drafting noise standards (dBA versus dBC and octave-band measures) and officials who need to explain to residents which level of government regulates generator emissions.
Limitations:
Describes Loudoun's own rules and experience; the county notes amendments to address tonal and low-frequency noise are anticipated in spring or summer 2027.
Cite as:
Loudoun County, Virginia (County Administrator). “Data Centers: Noise & Air Quality Concerns.” undated on page. https://www.loudoun.gov/6405/Noise-Air-Quality-Concerns
Environmental and Energy Study Institute (EESI), by Miguel Yañez-Barnuevo · March 23, 2026 · Web article
Part of EESI's data center impacts series, this article describes the sources of data center noise (cooling systems, diesel generators, fans), how noise is regulated at the local and state level through zoning ordinances, and EPA's dormant federal noise authority from the 1970s. Its key takeaways note that reliable sound level data is lacking and that most county and community noise ordinances were not written for continuous industrial noise.
“Data centers emit sounds from the humming of cooling systems, rumbling of diesel generators, and whirring of fans, which can be heard for hundreds of feet around them.” (Key Takeaways)
“Noise pollution is regulated at the local and state levels through zoning ordinances, but for a time in the 1970s, the Environmental Protection Agency oversaw noise pollution and conducted noise-control investigations.” (Key Takeaways)
Who should read it:
Council members and code enforcement staff who are hearing noise complaints and need background on why existing ordinances fall short.
Limitations:
Policy explainer from an environmental nonprofit; not original acoustic measurement.
Cite as:
Environmental and Energy Study Institute (EESI), by Miguel Yañez-Barnuevo. “Communities Are Raising Noise Pollution Concerns About Data Centers.” March 23, 2026. https://www.eesi.org/articles/view/communities-are-raising-noise-pollution-concernsabout-data-centers
Virginia Department of Environmental Quality · Issued Air Permits for Data Centers as of September 14, 2026 · State agency web page with permit list and guidance documents
Virginia DEQ lists the air permits it has issued for data centers (through its regional offices) and links two guidance documents: APG-576 on diesel engine-generator sets and APG-578 on sudden and reasonably unforeseeable events in planned electric outages. The page summarizes the April 2026 revision of APG-576, which sets presumptive Best Available Control Technology and performance testing requirements for emergency and non-emergency generator sets at data centers.
“The revised guidance document, APG-576 - Diesel Engine-Generator Set Procedure for Writing New and Modified Permits, became effective April 9, 2026.” (Air permit guidance documents)
“the presumptive BACT for both emergency and non-emergency gen-sets located as data centers is now based on the use of a selective catalytic reduction (SCR) system or equivalent for nitrogen oxides control” (Air permit guidance documents)
Who should read it:
Local officials who want to see what a state air agency now requires of data center generators, and residents' questions about who permits them.
Limitations:
Virginia-specific; individual permit documents were not opened in this session.
Cite as:
Virginia Department of Environmental Quality. “Issued Air Permits for Data Centers.” Issued Air Permits for Data Centers as of September 14, 2026. https://www.deq.virginia.gov/news-info/shortcuts/permits/air/issued-air-permits-for-data-centers
arXiv preprint (Yuelin Han, Zhifeng Wu, Pengfei Li, Adam Wierman, Shaolei Ren) · Submitted 9 Dec 2024; last revised 8 Jun 2026 (v4) · Academic paper
This paper (earlier versions were titled 'The Unpaid Toll') introduces a method to model criteria air pollutant emissions from data centers, including backup generators and power generation, and to estimate public health costs. The abstract projects national public health costs to 2028 and finds the burden is unevenly distributed, with the most affected counties facing per-household costs far above the national average. It proposes a 'health-informed computing' framework and recommends that energy reporting include public health impact.
“the growing demand for AI and computing technologies is projected to push the total annual public health burden of U.S. data centers up to more than $20 billion in 2028.” (Abstract)
“Although national-level impacts remain modest, data center health costs are unevenly distributed” (Abstract)
Who should read it:
Health departments and officials weighing cumulative air quality effects in counties with many data centers.
Limitations:
Preprint; results are modeled projections, and the title and content have been revised across four versions.
Cite as:
arXiv preprint (Yuelin Han, Zhifeng Wu, Pengfei Li, Adam Wierman, Shaolei Ren). “Health-Informed Computing: Estimating and Addressing the Public Health Impact of Data Centers.” Submitted 9 Dec 2024; last revised 8 Jun 2026 (v4). https://arxiv.org/abs/2412.06288
Environment America Research & Policy Center, Frontier Group and U.S. PIRG Education Fund (Quentin Good, Johanna Neumann, Abe Scarr) · May 7, 2026 · Web report
This report examines how data centers use banks of diesel backup generators, when they are permitted to run beyond true emergencies (scheduled outages, grid relief), and the difference in emissions between Tier 2 and Tier 4 generators. It discusses Virginia rule changes and argues battery storage is a cleaner alternative.
“These massive, energy-hungry facilities house dozens of industrial generators, many capable of producing between 1 and 3 megawatts (MW) of power each.” (Introduction)
“Tier 4 generators are significantly cleaner than those complying with Tier 2, cutting emissions of nitrogen oxides by 88% and particulate matter by 85%.” (Body)
“Until recently, emergency generators (dirtier Tier 2 generators) in Virginia were not allowed to run during planned power outages.” (Body)
Who should read it:
Officials negotiating conditions on generator tier, run hours or battery backup.
Limitations:
Published by environmental advocacy organizations.
Cite as:
Environment America Research & Policy Center, Frontier Group and U.S. PIRG Education Fund (Quentin Good, Johanna Neumann, Abe Scarr). “False emergencies, real pollution: Diesel generator use at data centers.” May 7, 2026. https://environmentamerica.org/center/resources/false-emergencies-real-pollution/
This nonprofit newsroom article reports on how diesel backup generators at data centers are permitted, with a focus on Texas (TCEQ) and federal limits on non-emergency run hours, including hours that can be used for demand response. It covers noise and pollution concerns and industry interest in gas generators as an alternative.
“But for non-emergencies, like machine testing, diesel generators are federally limited to 100 hours.” (Body)
“Fifty of those 100 hours can be used for demand response programs” (Body)
Who should read it:
Officials who need a plain explanation of generator run-hour limits and demand response.
Limitations:
Journalism rather than research. The article's actual headline uses an em dash; it is rendered here with a colon.
Cite as:
Inside Climate News (Arcelia Martin). “Data Centers’ Use of Diesel Generators for Backup Power Is Commonplace: and Problematic.” November 12, 2025. https://insideclimatenews.org/news/12112025/data-center-diesel-generators-noise-pollution/
Loudoun County, Virginia (County Administrator) · undated on page (references the adopted FY 2027 budget and tax year 2026 rates) · County web page
Loudoun County explains how data center revenue, mainly business personal property tax on computer equipment plus real estate tax, has shaped its budget. Tabs cover Tax Rates, Data Center Revenue, Data Centers and Taxes, and Stabilizing Budget Growth. The county lists what the revenue has funded (schools, public safety, libraries, parks, capital projects) and warns against overdependence on one revenue source, projecting that personal property tax growth from data centers will plateau.
“Over the past ten years, the Board of Supervisors has generally lowered the real property tax rate every year from $1.145 in tax year 2016 to $0.805 cents per $100 in assessed value for tax year 2026.” (Tax Rates tab)
“the general personal property tax rate, which applies to property in data centers, remains the same as the current rate for tax years 2026 and 2027: $4.15 per $100 in assessed value.” (Tax Rates tab)
Who should read it:
Finance directors and budget committees modeling how data center revenue could be used and what happens when it stops growing.
Limitations:
Reflects one county with the world's largest concentration of data centers and Virginia's tax structure; results will differ where equipment is exempted or abated.
Cite as:
Loudoun County, Virginia (County Administrator). “Data Centers: Tax Revenues & County Budget.” undated on page (references the adopted FY 2027 budget and tax year 2026 rates). https://www.loudoun.gov/6409/Tax-Revenues-County-Budget
This county page addresses aesthetics, residential property values and workforce effects of data centers. On property values it explains why a measurable effect is hard to detect in a tight housing market and that assessments have not shown discounts near data centers. On jobs it gives a direct and indirect employment estimate and notes each data center needs relatively few staff once operating.
“Assessing the impact of data centers on residential property values is complicated.” (Land values section)
“no impacts to housing prices due to proximity to a data center are apparent at this time.” (Land values section)
“Data centers are responsible for the creation of more than 15,000 jobs, directly and indirectly, in Loudoun County.” (Workforce Development and Jobs section)
Who should read it:
Assessors and officials fielding resident questions about home values near proposed sites.
Limitations:
The county itself says the housing market is tight and that effects could appear if demand waned; this is an assessor's observation, not a statistical study.
Cite as:
Loudoun County, Virginia (County Administrator). “Data Centers: Appearance, Land Values & Workforce.” undated on page. https://www.loudoun.gov/6407/Appearance-Land-Values-Workforce
National Bureau of Economic Research (Fernando E. Alvarez, David Argente, Joyce Chow, Diana Van Patten) · Issue date May 2026; revision date September 2026 · Academic working paper
The authors link a facility-level panel of U.S. data centers to county measures of employment, establishments, payroll, income, house prices, electricity prices and public-supply water withdrawals. To handle the fact that developers choose sites, they build an instrument from historical fiber-route geography (railroad corridors and backbone access points). They report effects after the modern expansion began, and dynamic long-difference estimates over time.
“the estimates show positive effects on total employment, construction employment, establishments, payroll, tax returns, adjusted gross income, and wages.” (Abstract)
“sustained increases in house prices and public-supply withdrawals, and positive effects on electricity prices that emerge later.” (Abstract)
Who should read it:
Economic development directors and finance staff who want peer-level causal evidence rather than industry impact studies.
Limitations:
Working paper, not yet peer reviewed; only the abstract page was read.
Cite as:
National Bureau of Economic Research (Fernando E. Alvarez, David Argente, Joyce Chow, Diana Van Patten). “Data Centers and Local Economies in the Age of AI: A Shift: Share Approach (NBER Working Paper 35194).” Issue date May 2026; revision date September 2026. https://www.nber.org/papers/w35194
Brookings Institution (Dany Bahar and Greg Wright) · Originally published May 4, 2026; updated August 10, 2026 · Research brief
This brief summarizes the authors' paper 'Data Centers and Local Labor Markets,' which compares labor markets that received their first large data center with control markets. It reports effects on data processing and telecommunications employment, wages and home prices over the first decade, and finds effects differ by facility type. An editor's note says earlier findings on employment, wages and home prices were revised with an expanded sample.
“Labor markets that receive their first large data center see employment rise in the data processing sector by 56% over the first decade of operations.” (Findings)
“Wages were unchanged, and we find a modest increase in home prices of 2%-5%.” (Findings)
“At a typical treated county, these estimates imply roughly 100-200 jobs, depending on facility type.” (Findings)
Who should read it:
Officials evaluating job claims in a developer's application or an incentive request.
Limitations:
The editor's note says results were revised and research is continuing; the full paper is 'available from the authors.'
Cite as:
Brookings Institution (Dany Bahar and Greg Wright). “New evidence on data center employment effects.” Originally published May 4, 2026; updated August 10, 2026. https://www.brookings.edu/articles/new-evidence-on-data-center-employment-effects/
Brookings Institution (Daniel Goetzel, Mark Muro, Shriya Methkupally) · February 5, 2026 · Web article
The authors argue that the standard model of fast, opaque data center dealmaking delivers short-term construction jobs and revenue but little lasting local economic benefit. They say competition for sites, grid access and permits gives regions leverage, and cite early examples of negotiated co-investments that anchor regional tech ecosystems.
“Regions should treat data center negotiations not as isolated real estate transactions but as ecosystem-shaping moments that trade infrastructure access for commitments to advance local innovation, talent, and industry strengths.” (Summary bullets)
Who should read it:
Economic development officials preparing to negotiate development or community benefit agreements.
Limitations:
Policy commentary; examples are illustrative.
Cite as:
Brookings Institution (Daniel Goetzel, Mark Muro, Shriya Methkupally). “Turning the data center boom into long-term, local prosperity.” February 5, 2026. https://www.brookings.edu/articles/turning-the-data-center-boom-into-long-term-local-prosperity/
World Resources Institute (Carla D. Walker and Ian Goldsmith) · February 17, 2026 · Web explainer
WRI's explainer walks through seven impact areas: energy demand and power bills, local water supplies, air pollution and climate, noise, competition for land, pressure on marginalized communities, and economic trade-offs. Each section pairs the impact with governance responses (for example dedicated rate classes, water-use disclosure, battery storage instead of diesel, setbacks and noise monitoring, agricultural zoning, community benefits agreements). It ends with a 'Summary of Governance Tools' table and examples of local action such as Jerome Township, Ohio's moratorium, Prince George's County, Maryland's task force and Lancaster, Pennsylvania's community benefits agreement.
“Two-thirds of all data centers built or in development since 2022 are located in water-stressed areas like southern Arizona, the Colorado River Basin and Texas.”
“In 2024, the average data center site covered about 224 acres or 0.35 square miles”
“A review of more than 1,200 U.S. data centers found that even the largest employ fewer than 150 permanent workers, and sometimes as few as 25.”
“A review of 31 Virginia municipalities with existing or proposed data centers found that 25 (80%) had NDAs in place.” (What Governments Can Do section)
Who should read it:
Any local official who needs one readable overview of the full range of community impacts and a menu of tools; a good first read before commissioning local studies.
Limitations:
Explainer that synthesizes other studies rather than presenting new data; WRI is an environmental research organization.
Cite as:
World Resources Institute (Carla D. Walker and Ian Goldsmith). “From Energy Use to Air Quality, the Many Ways Data Centers Affect US Communities (page title: 7 Ways Data Centers Affect US Communities).” February 17, 2026. https://www.wri.org/insights/us-data-center-growth-impacts
Wisconsin Farm Bureau Federation · undated on page (cites WEDC certifications 'as of February 2025') · Issue backgrounder
Written for farmers facing a nearby data center proposal, this backgrounder covers land requirements, energy demand and who pays for power infrastructure, cooling and water use (including rural high-capacity wells), stormwater runoff, 24/7 lighting and noise, utility easements, jobs, and Wisconsin's sales and use tax exemption for certified data centers. It closes with considerations for farmers before development.
“Data centers need a minimum of 40 acres for a small center and upwards of 1000 acres for a large center.” (How Much Land Do Data Centers Require?)
“rural locations may require multiple high-capacity wells running on an ongoing basis.” (Groundwater, Wells, and Rural Water Supply Concerns)
“in the long term they support a rather limited number of jobs for the size of the investment.” (Noise, Lighting, and Quality of Life section)
Who should read it:
Rural township and county officials, and farm neighbors, weighing rezoning of agricultural land.
Cite as:
Wisconsin Farm Bureau Federation. “Data Center Impacts on Agriculture.” undated on page (cites WEDC certifications 'as of February 2025'). https://wfbf.com/policy/current-issues/state-agricultural-issues/data-center-impacts-on-agriculture/
About this library
Why we put this together.
When a data center shows up on an agenda, the questions come fast, and the answers are scattered across federal reports, utility filings, state bills and local ordinances. We wanted one place a council member, a planner or a neighbor could start. We read what we list, and if we only got through part of a long report, we tell you which part. If something sat behind a paywall or the link was dead, we left it out.
Each entry includes
Who published it, when, and a link.
A short description of what it covers.
Quotes taken directly from the source.
Who should read it.
Its limits, including who paid for it.
A citation for your staff report.
We update the library as new information becomes available. No one paid to be listed, and government, industry and advocacy sources sit side by side, each one labeled.
Is a data center coming to your community?
The library is free. If you want an independent read before the first hearing, we can help.