Data centre noise harms health of nearby residents, US survey finds
A joint survey by two US nonprofits has found that residents living near data centres and cryptomining facilities report widespread noise-related health harms, with a large majority able to hear facility noise inside their homes. The research, conducted by Quiet Communities and Science for Georgia and approved by the University of Georgia's Institutional Review Board, is the first nationally scoped survey to quantify resident experiences of operational noise from these facilities.
The survey drew 103 responses across 11 US states. Among those who could hear the noise indoors, 71% said it interfered with their ability to rest and relax, 65% reported disturbed sleep, and 54% found it difficult to concentrate. Broader concerns were also pronounced: 95% of respondents expressed worry about mental health impacts, 87% about physical health, and 90% about falling property values. Sixty-seven per cent said they had considered leaving their homes because of the noise.
What the noise looks like in practice
Both data centres and cryptomining facilities depend on industrial cooling equipment, HVAC systems and diesel backup generators that run continuously, round the clock. Recorded sound levels at these sites can exceed 80 dBA, a level the researchers compare to a vacuum cleaner or factory floor. Unlike a transient noise event, the operational hum persists for months or years, exposing nearby residents to what many in the survey described as a constant low-frequency drone.
One finding with direct regulatory implications: more than 76% of respondents who lived over half a mile from a facility could still hear the noise inside their homes. Current zoning buffers in many US jurisdictions sit at around 500 feet, roughly a tenth of a mile. Jamie Banks, founder and president of Quiet Communities, said the survey "begins to fill the gap" in documented resident experience. "The results suggest these impacts are widespread enough that they deserve serious attention from policymakers and the broader industry," Banks added.
Prior epidemiological research cited by the authors linked long-term environmental noise exposure to more than 45,000 new cases of heart disease and 12,000 premature deaths in 2020 alone, figures drawing on established noise-health literature rather than this survey specifically.
Regulatory and market context
The survey lands at a moment of extraordinary growth in US data centre construction. The release notes that as of April 2026, roughly 42% of Americans live within five miles of an existing or planned facility, with more than 3,000 sites currently operating and approximately 1,500 additional projects in the pipeline. Demand is being driven primarily by AI workloads; separately, some former cryptomining operators are repurposing or co-locating infrastructure to serve AI compute demand, blurring the regulatory distinction between the two facility types.
The US has no federal noise standard specifically governing data centres. Regulation is fragmented across state environmental agencies and local zoning ordinances, leaving communities to rely on general nuisance law or negotiate site-specific conditions during planning approval. The European Union's position is only marginally more structured: the Environmental Noise Directive requires member states to map major industrial noise sources and produce action plans, but thresholds for facilities comparable to hyperscale data centres remain inconsistently applied.
Dr Amy Sharma, executive director of Science for Georgia, argued that the pace of development has outrun independent evidence on community impact. "Communities are being asked to host facilities without the benefit of robust, independent research on what living next to them actually means over the long term," she said.
The authors acknowledge the survey's limitations: a sample of 103 self-selected respondents recruited partly through community advocacy networks introduces selection bias, and the study does not include objective noise measurements at respondent properties. Policymakers and industry operators will likely cite those methodological constraints. Nonetheless, the IRB-approved framing gives the findings more institutional weight than a purely advocacy-led report, and the researchers say the work is intended to prompt further independent study and to inform the development of more robust noise ordinances at state and local level.