Turns out ‘there is something to it…’
Electrical Gluttony. Water Consumption. Noise Pollution. Ground Vibration. Powerline Harmonics. Incessant Toxic Badness.
10+part sourced Series deeply explores the Many Serious Ill-Effects of Datacenters.
This is not us merely saying these things are or may be so; instead, every statement is sourced by scientific studies, published experiences, harmed victims. Intelligence requires learning from own mistakes. Wisdom involves learning from others’ mistakes.
“A 2026 power-engineering study found that hyperscale data centers behave not merely as electrical loads, but as vast power-electronics systems whose rapidly fluctuating server demands can excite grid oscillations and propagate disturbances across interconnected networks. Far from a fringe “dirty electricity” claim, the research suggests data centers may interact with the power grid in ways fundamentally different from traditional industrial facilities, raising serious new concerns about resonance, stability, reliability.”
1. Direct Emissions: no alpha / beta / gamma evidence
Modern servers, GPUs, routers, fiber optics, UPS units, chillers, and transformers do not contain mechanisms that normally emit alpha, beta, or gamma radiation. They emit heat, low-frequency electric/magnetic fields, and RF/optical signals — mostly non-ionizing.
NRC’s basic ionizing radiation categories are alpha, beta, gamma/x-ray, and neutron radiation; ordinary electronic data-center equipment does not fall into those source categories unless it contains specific radioactive sources or x-ray-producing high-voltage vacuum devices, which normal server farms do not.
Source: NRC radiation basics:
https://www.nrc.gov/about-nrc/radiation/health-effects/radiation-basics
2 Radon in data centers: possible, not unique
Radon comes from uranium/radium decay in soil and rock. EPA says it can enter buildings through foundations, cracks, joints, pipes, and water. That applies to homes, offices, schools, warehouses — and data centers.
EPA radon basics:
https://www.epa.gov/radiation/radionuclide-basics-radon
There is a workplace literature on radon but no strong body of papers saying “data centers have elevated radon because they are datacenters.” “A data center built on radon-prone geology could have radon issues like any large building, especially if slab/foundation design, ventilation pressure, or sub-slab mitigation is poor.”
NIOSH/PMC workplace radon review:
https://pmc.ncbi.nlm.nih.gov/articles/PMC5751755/
3 Vibration increasing radon release: yes, some evidence
This is the most interesting part of the ‘radioactivity hypothesis’.
A 1998 Health Physics paper studied traffic vibration near roads and railways and found soil-gas radon generally increased with vibration. It proposed a “pump effect” where mechanical vibration moves soil gas upward. It reported higher increases near motor traffic and rail, with effects decaying over distance.
Study: “The Influence of Traffic Vibrations on the Radon Potential”
https://journals.lww.com/health-physics/abstract/1998/02000/the_influence_of_traffic_vibrations_on_the_radon.8.aspx
A 2018 lab study on uranium-bearing similar materials found low-frequency vibration affected radon exhalation, with increases up to certain frequency ranges.
Study: “Laboratory Experimental Laws for the Radon Exhalation of Similar Uranium Samples with Low-Frequency Vibrations”
https://www.mdpi.com/2071-1050/10/8/2937
A Peru soil-radon paper also cites vibration/radon work and discusses radon emanation/exhalation in relation to traffic/ground vibration.
https://www.scielo.org.co/scielo.php?pid=S1794-61902019000300171&script=sci_arttext
Relevance to data centers:
Operational vibration from fans/chillers/generators is probably much less important than:
- construction blasting
- pile driving
- heavy truck traffic
- excavation
- quarrying/mining nearby
- rail lines
- seismic/fault-zone activity
But if a data center has huge mechanical plant or continuous generator testing on radon-prone fractured bedrock, measuring soil-gas/indoor radon before/after would be reasonable.
4 Piezoelectric / rock-stress radiation: mostly EM, not radioactive particles
There is legitimate research on stressed/fractured rocks producing electrical currents, surface potentials, infrared, visible light, RF, and other electromagnetic phenomena.
USGS has work on earthquake lights and stress-activated charge carriers in rocks.
USGS:
https://www.usgs.gov/publications/earthquake-lights-and-stress-activation-positive-hole-charge-carriers-rocks
There are also lab studies on visible-light and x-ray emissions during mineral/rock fracture, especially involving quartz/fracture surfaces.
MDPI Minerals paper:
https://www.mdpi.com/2075-163X/12/6/778
But that is not the same as ongoing alpha/beta/gamma radioactivity from a data center. Rock fracture can produce transient electrical/light/x-ray phenomena in controlled or extreme conditions; it does not imply ordinary facility vibration creates radioactive emissions.
There are controversial “piezonuclear” claims of neutron emission from fracturing rocks.
Claim example:
https://journals.sagepub.com/doi/10.3233/SFC-2011-0120
5 Groundwater / radioactive water: plausible only if local geology has radionuclides
Groundwater can naturally contain radon, uranium, radium, gross alpha/beta activity, etc. USGS explicitly notes radionuclides occur in groundwater and are controlled by geology/geochemistry.
USGS radionuclides:
https://www.usgs.gov/mission-areas/water-resources/science/radionuclides
USGS overview:
https://www.usgs.gov/publications/radionuclides-surface-water-and-groundwater
EPA regulates radionuclides in drinking water, including gross alpha, beta/photon emitters, radium, and uranium.
EPA radionuclides rule:
https://www.epa.gov/dwreginfo/radionuclides-rule
Data centers can use significant water for cooling. Scholarly work confirms major data-center water footprints and groundwater/surface-water implications.
ORNL / Environmental footprint of U.S. data centers:
https://impact.ornl.gov/en/publications/the-environmental-footprint-of-data-centers-in-the-united-states
Nature npj Clean Water:
https://www.nature.com/articles/s41545-021-00101-w
But using water does not make it radioactive. The risk would be:
- pumping naturally radon/radium/uranium-bearing groundwater,
- aerating radon-bearing water,
- concentrating radionuclides in treatment residuals,
- discharging water in a way that mobilizes existing NORM.
6 Power infrastructure: EMF yes, ionizing radiation no
Data centers often require substations, transmission lines, transformers, and high-current distribution. These create low-frequency electric/magnetic fields. That’s non-ionizing EMF, not alpha/beta/gamma; nonetheless, studies have shown ill-effects.
NIEHS says some studies show a possible weak association between EMF field strength and childhood leukemia, but findings are not definitive.
NIEHS EMF:
https://www.niehs.nih.gov/health/topics/agents/emf
EPA notes WHO/IARC classified extremely low-frequency EMF as “possibly carcinogenic” based on limited evidence for childhood leukemia.
EPA power-line EMF:
https://www.epa.gov/radtown/electric-and-magnetic-fields-power-lines
NCI summarizes that epidemiologic studies show increased childhood leukemia risk above about 0.3–0.4 µT daily average exposure, but mechanisms have not been identified and experimental support is lacking.
NCI EMF fact sheet:
https://www.cancer.gov/about-cancer/causes-prevention/risk/radiation/electromagnetic-fields-fact-sheet
Data centers are not known direct sources of alpha, beta, or gamma radiation; however, a data center can intersect with natural radioactivity through geology, radon, groundwater, construction vibration, and upstream power generation. The only credible way to prove a local effect is site-specific baseline and post-construction monitoring.