AI Data Centers and Low-Frequency Vibration: Health Effects Explained

Large data center facility located in a residential community
AI hyperscale data centers produce continuous low-frequency hums and infrasound from cooling systems and generators. This post reviews community impacts, documented complaints, and the current scientific evidence on health effects.

Low-frequency vibration is an emerging concern tied to the rapid growth of AI data centers, spanning community impacts from external noise/infrasound, internal equipment reliability challenges, and related power-system effects.  

Infrasound (sound waves typically below 20 Hz, often inaudible) has been studied for potential health effects for decades. Research spans animal experiments, controlled human laboratory trials, and observational/epidemiological studies. Findings are highly dependent on intensity (sound pressure level in dB), frequency, duration, and whether the exposure includes audible components. High-intensity exposures show clearer acute effects; environmental or typical industrial levels (e.g., near wind turbines, industrial fans, or data centers) yield more mixed or limited evidence of harm.

Ad: Remove Your Information from Databases
The Know Your H20 Team uses
( Incogni)

External Impacts: Community Noise, Infrasound, and Felt Vibrations


AI hyperscale facilities (and related high-density computing sites) produce continuous low-frequency hums and infrasound (typically below 20 Hz) primarily from cooling systems (large fans, HVAC, cooling towers), diesel/gas backup generators, and on-site turbines. These frequencies attenuate poorly with distance, penetrate buildings more readily than higher-frequency sound, and are often perceived as pressure or physical vibration rather than audible noise. Standard A-weighted (dBA) measurements undervalue them, so complaints can exceed conventional noise ordinance readings.


Residents near facilities have reported symptoms including vertigo, nausea, migraines, insomnia, tinnitus, dizziness, anxiety, and a sense of pressure. Documented examples include:


  • A Bitcoin mining site (300-megawatt) in Granbury, Texas (near homes), linked to similar complaints from cooling and power systems. (Article)

  • xAI's Southaven, Mississippi, power plant (supporting nearby data centers with dozens of natural gas turbines), which has drawn lawsuits alleging near-constant noise, low-frequency rumbling, and vibrations felt inside homes, disrupting sleep and daily life. Residents have described jet-engine-like sounds combined with felt vibration; mitigation efforts (including sound barriers) have been criticized as insufficient by some.


Health effects of pure infrasound at typical industrial levels remain an open debate in the literature. Some studies show limited to no clear physiological impacts in controlled settings and others note possible or suggested stresses. Chronic audible low-frequency industrial noise is more clearly associated with sleep disruption and related issues.


High-Intensity Acute Effects (Well-Established)

Animal and limited human data show biological effects mainly above ~100 dB, with clearer damage above ~130–140 dB:

  • Ear damage, tympanic membrane issues, hearing threshold shifts, or vestibular symptoms (dizziness, nausea) at very high levels.

  • Cardiovascular changes (e.g., blood pressure, heart rate), neurochemical shifts, or reduced performance in animal models at high intensities (often 120+ dB for hours).

  • Aural pain or eardrum rupture possible at extreme levels.


These levels far exceed typical environmental exposures. The UK's Advisory Group on Non-ionizing Radiation (AGNIR/Public Health England RCE-14, 2010) concluded there is no consistent evidence of physiological or behavioral effects from acute human exposure at lower levels, with sparse long-term data and no established ill-effects.


Controlled Human Experimental Studies (Low-to-Moderate Levels)


Modern randomized trials often find null or minimal effects for inaudible or near-threshold infrasound:

  • 2021 Scientific Reports pilot RCT (N=38 healthy adults): 6 Hz at 80–90 dB SPL for 28 nights (bedroom devices) vs. sham. No significant effects on psychiatric symptoms, sleep quality, stress, sound sensitivity, or cognition. Exploratory findings included a trend toward more physical weakness and decreases in regional grey matter volume (cerebellum and angular gyrus), without clear brain-behavior links. Conclusion: no broad effects on mental health, cognition, or global brain structure in healthy people.

  • 2023 Environmental Health Perspectives double-blind crossover (N=37 noise-sensitive healthy adults): 72 hours of simulated wind turbine infrasound (1.6–20 Hz, ~90 dB peak) vs. sham vs. audible traffic noise control. No impact on primary outcome (wake after sleep onset), other sleep physiology, cardiovascular measures, neurobehavioral performance, or "wind turbine syndrome" symptoms. Traffic noise (positive control) did disrupt sleep. Participants could not distinguish infrasound from sham. Did not support infrasound as a cause of wind turbine syndrome symptoms. (Note: "sham" refers to a placebo or quiet control condition.)

  • 2026 Frontiers in Behavioral Neuroscience study (N=36): ~18 Hz infrasound (75–78 dB) mixed with calming or unsettling music (2×2 design). Infrasound (undetectable above chance, no expectancy effects) was associated with elevated salivary cortisol and higher self-reported irritability, disinterest, and sadness appraisal. Suggests possible subtle stress/aversive responding even without conscious detection.


Older or higher-intensity lab work sometimes reported temporary annoyance, vibration sensations, or mild physiological changes near or above hearing thresholds (hearing thresholds for infrasound are high—often 80–110+ dB depending on frequency).

Epidemiological and Observational Evidence


A 2016 systematic review of observational studies found associations between low-frequency noise (LFN, including infrasound components) and self-reported annoyance, sleep-related problems, concentration difficulties, and headache in adults near relevant sources. Pooled prevalence of high subjective annoyance attributed to LFN was ~10%. Evidence for chronic medical conditions was very limited; studies had methodological shortcomings (e.g., exposure assessment, confounding).

Narrative reviews (e.g., 2020–2023 Polish Medycyna Pracy series by Pawlaczyk-Łuszczyńska et al., covering experimental and epidemiological literature up to ~2022) emphasize annoyance as the primary outcome at everyday levels, with possible links to sleep quality and concentration. Cardiovascular effects and other outcomes show inconsistent or limited support. Special attention to wind turbines finds that total sound level and amplitude modulation better explain annoyance than the infrasound component alone.

Updates on wind turbine sound (e.g., 2021 IJERPH) conclude that infrasound well below the hearing threshold is unlikely to produce effects beyond those of audible sound; low-frequency content behaves like other noise regarding annoyance and potential sleep impacts.

A more recent narrative review (Noise & Health, ~2025) notes possible influences on cardiovascular, vestibular, and neural pathways, with outcomes depending strongly on frequency, intensity, and duration. It highlights research gaps in mechanisms and calls for better surveillance, while acknowledging debate around claims like "wind turbine syndrome."

Overall Consensus and Limitations

  • Clear risks exist mainly at very high intensities (rarely encountered environmentally).

  • At typical environmental/industrial levels (often well below or near hearing thresholds for pure infrasound), controlled studies frequently show no robust effects on sleep, cognition, or objective physiology. Self-reported annoyance and symptoms occur and can be distressing, potentially mediated by audible noise, expectation, or individual sensitivity.

  • Confounding is common: low-frequency noise rarely occurs in isolation from audible components; vibration (structure-borne) and psychological factors (nocebo) complicate attribution.

  • Evidence quality is often limited by small samples, short durations, poor exposure characterization, and reliance on self-report.

  • Research gaps include long-term low-level chronic exposure, precise dose-response for subtle effects (e.g., stress markers), sensitive subgroups, and mechanistic pathways.


In the context of AI data centers or similar industrial sources, complaints about felt vibration or pressure often involve a mix of low-frequency audible noise, infrasound, and structure-borne vibration. Standard dBA metrics under-weight low frequencies, so specialized measurements (e.g., dBC, narrowband analysis, or vibration sensors) are more relevant. Current high-quality experimental data do not strongly support major health impacts from pure infrasound at typical community levels, but ongoing research, better monitoring, and mitigation of overall noise remain warranted.

Concerned about your drinking water quality? Environmental noise and infrastructure changes near your home can raise broader questions about local water safety. Our Drinking Water Guide for Well Water & City Water by Brian Oram provides clear, science-based guidance on testing, understanding results, and protecting your household water supply. Order your copy today →

 

Citations

Baliatsas, C., van Kamp, I., van Poll, R., & Yzermans, J. (2016). Health effects from low-frequency noise and infrasound in the general population: Is it time to listen? A systematic review of observational studies. Science of the Total Environment, 557–558, 163–169. https://www.sciencedirect.com/science/article/abs/pii/S0048969716304338?via%3Dihub

Dommes, E., et al. (2021). A longitudinal, randomized experimental pilot study to investigate the effects of airborne infrasound on human mental health, cognition, and brain structure. Scientific Reports, 11, Article 3190. https://www.nature.com/articles/s41598-021-82203-6

Independent Advisory Group on Non-ionizing Radiation (AGNIR). (2010). Health effects of exposure to ultrasound and infrasound (RCE-14). Public Health England / Health Protection Agency. https://www.gov.uk/government/publications/ultrasound-and-infrasound-hpa-response-to-agnir-report-rce-14/ultrasound-and-infrasound-hpa-response-to-agnir-report-rce-14

Maijala, P., et al. (or relevant authors). (2021). Health effects related to wind turbine sound: An update. International Journal of Environmental Research and Public Health, 18(17), Article 9133. https://www.mdpi.com/1660-4601/18/17/9133

Marshall, N. S., et al. (2023). The health effects of 72 hours of simulated wind turbine infrasound: A double-blind randomized crossover study in noise-sensitive, healthy adults. Environmental Health Perspectives, 131(3), 037012.  https://pubmed.ncbi.nlm.nih.gov/36946580/

Pawlaczyk-Łuszczyńska, M., Dudarewicz, A., Myshchenko, I., & Bortkiewicz, A. (2023). Impact of infrasound and low frequency noise on human health and well-being. Part II: Review of epidemiological studies. Medycyna Pracy, 74(5), 409–423. https://medpr.imp.lodz.pl/Wplyw-infradzwiekow-i-halasu-niskoczestotliwosciowego-na-zdrowie-i-samopoczucie-czlowieka,172194,0,2.html

Recent narrative review authors. (2025). Infrasound in biology and medicine: Insights into mechanisms, health outcomes and research perspectives – A narrative review. Noise & Health, 27(129), 676–691. https://www.ovid.com/jnls/nohe/fulltext/10.4103/nah.nah_136_25~infrasound-in-biology-and-medicine-insights-into-mechanisms

2026 Frontiers study authors. (2026). Infrasound exposure is linked to aversive responding, negative appraisal, and elevated salivary cortisol in humans. Frontiers in Behavioral Neuroscience. https://www.frontiersin.org/journals/behavioral-neuroscience/articles/10.3389/fnbeh.2026.1729876/full

Additional Reading Suggestions

Suicidal Empathy (Amazon Affiliate Program)

The Nerd Reich: Silicon Valley Fascism and the War on Democracy (Amazon Affiliate Program)

If you are going to have your drinking water tested using informational water testing, please consider using the NTL Mail Order Testing Service (Well Water / City Water).

Other Websites

Know Your H20
B.F. Environmental Consultants
Carbon County Groundwater Guardians
Keystone Clean Water Team (Donate)

Affiliate Disclosure: This article contains affiliate links. If you purchase a product through one of these links, we may earn a small commission at no additional cost to you. We only recommend products we believe in. This helps support our mission to educate consumers about water quality.

Share