Summary: The Rapid Growth of AI Data Centers is Degrading Power Quality and Increasing Harmonics and Electrical Pollution in Communities Nationwide

In addition to the air, water, soil, and other environmental pollution associated with the rapid expansion of AI data centers, this infrastructure is contributing to measurable harmonic distortion, degrading power quality for millions of U.S. households, and thus, increasing electromagnetic field (EMF) exposures in homes.

Data centers rely heavily on electrical components that generate significant harmonics, as detailed in a 2026 publication by Sheng et al. in the journal Energies which states, “this harmonic distortion not only degrades power quality within the facility (e.g., causing overheating of internal transformers) but also propagates to the distribution network, affecting neighboring users (e.g., triggering malfunctions in precision manufacturing equipment. Voltage sags and flicker are additional prominent issues.”

A blockbuster Bloomberg report analysing measurements from 770,000 residential sensors across the US found that data center harmonics can extend for miles outside the facility. Three-fourths of the most highly distorted power readings occurred within 50 miles of large data center activity. More than half of the households experiencing the worst distortions were located within 20 miles of significant data center activity.

What Are Harmonics and Dirty Electricity?

Electricity is delivered to homes as alternating current (AC) at 60 Hz in North America and 50 Hz in Europe and most of the world. Ideally, this electricity follows a smooth sinusoidal waveform.

Harmonics occur when the normal 60 or 50 Hz wave used to deliver electricity is distorted by additional frequencies, disrupting the smooth flow of electricity and causing erratic spikes and dips in voltage that can damage and degrade equipment, including household appliances and electronics. 

Distortions in voltage and current, collectively known as poor “power quality” by the electric utilities, can overheat wiring, trigger electrical faults or arcing, and even increase the risk of sparks and electrical fires.  Further, harmonic distortion can also alter the electrical environment inside homes, affecting human health, as higher-frequency currents can travel along interior household wiring and increase EMF exposure within living spaces.

• THDB: Total harmonic distortion of the magnetic field (THDB) is the measure of how distorted the magnetic field  is compared to a clean, smooth electrical signal. 
• THD is total harmonic distortion, a measure of how much an electrical signal (voltage or current) is distorted compared to a clean, smooth waveform.

Because magnetic fields are created by electrical currents, distorted electricity (high THD) leads to distorted magnetic fields (high THDB), meaning the issue isn’t just inside the wiring, but also extends into the surrounding environment, meaning into the air of rooms where people live and work. Transients, harmonics, and related electrical distortions are also often referred to as “dirty electricity.”  

Evidence of Harmonic Distortion Near Data Centers

Bloomberg’s analysis of ~770,000 residential sensors found that more than three-quarters of the highest distorted power readings occurred within 50 miles of data centers, and more than half of the homes experiencing the worst distortions were within 20 miles of major data centers. Some areas recorded total harmonic distortion (THD) levels above 8%, the threshold engineers say exceeds accepted limits. In Northern Virginia’s “data center alley,” sensors recorded distortions as high as 12.9% THD, while counties far from data centers averaged below 3%.

It is especially important to understand the power system impact of AI “because it is such a big hammer” on the grid, stated Hasala Dharmawardena, a senior member of the Institute of Electrical and Electronics Engineers, in the DataCenter Dynamics article “AI data centers causing “distortions” in US power grid.”

The study, “Monitoring Harmonic ELF Magnetic Fields in Data Centers with Sensor-Based Instrumentation” published in IEEE Transactions on Instrumentation and Measurement documented the results of a one-month assessment of the extremely low-frequency (ELF) magnetic field as well as harmonic distortions in a data center. The researchers found that typical magnetic field levels were around 5 milligauss (mG) (0.5 μT), with a maximum below 10 mG (1 μT), while distortion levels were consistently high, with a median total harmonic distortion of the magnetic field (THDB) around 36% and occasional values exceeding 300%. Daily and weekly patterns showed that distortion was much higher during busy working hours, about twice as high as at night and on weekends.  

Health Implications of Altered Electrical Environments

Because the human body is electrically conductive and relies on electrical signaling for physiological processes, external electromagnetic field exposure can impact biology. Although studies funded by the electrical industry more often report no significant effects, and industry-sponsored reviews downplay research showing adverse outcomes from non-ionizing electromagnetic fields, the majority of peer-reviewed studies report measurable biological effects from non-ionizing EMF exposure.  

Both extremely low-frequency (ELF) magnetic fields and wireless (cell tower) radiofrequency EMF were classified by the World Health Organization’s International Agency for Research on Cancer (IARC) as a possible human carcinogen and the last few decades of science have continued to report associations between non-ionizing EMF and increased cancers. 

Hyperscale data centers require massive amounts of electricity, often necessitating significant expansion of the surrounding power grid. This includes running new high-voltage transmission lines through nearby communities, which can further increase residential EMF exposure. Hundreds of scientists recommend reducing EMF exposure due to the scientific evidence linking exposure to adverse health effects. Notably, ELF magnetic fields are associated with increased risk for childhood leukemia at levels thousands of times lower than industry standards. Kaiser Permanente studies report that prenatal ELF-EMF exposure was associated with increased miscarriage risk and also ADHD, obesity, and asthma. 

Watch a video of Dr. De Kun-Li presenting on his team’s research on EMF magnetic fields and health below.

Dr. Henry Lai, Professor Emeritus at the University of Washington and Editor Emeritus of the journal Electromagnetic Biology and Medicine, has compiled summaries of the research on the biological effects of EMF exposure, concluding that the majority of research has found that exposure produces oxidative effects and genetic, neurological, reproductive, and developmental impacts. This information is posted on Dr. Joel Moskowitz website SaferEMR.com. 

Dr. Samuel Milham, an epidemiologist formerly at the Washington State Department of Health, published the book Dirty Electricity: Electrification and the Diseases of Civilization as well as several scientific papers on how EMF exposure, including high-frequency voltage transients, is a significant environmental stressor and contributor to numerous diseases. His study of teachers in a California school found that elevated exposure to high-frequency voltage transients on electrical wiring was associated with increased cancer incidence among staff. In another publication “Evidence that dirty electricity is causing the worldwide epidemics of obesity and diabetes,” Dr. Milham hypothesized that high-frequency voltage transients (harmonics or “dirty electricity) contribute to chronic diseases such as obesity and diabetes, documenting how diesel generators (known sources of electrical transients) are widely used to electrify certain pacific islands and remote regions, several of which have reported high rates of metabolic disease. 

Managing Harmonics, EMFs and Protecting Communities 

Power quality issues must be addressed proactively. Sheng et al. 2025 recommends harmonic distortion and other power quality issues must be actively managed through filters and higher quality equipment, stricter grid standards, and smarter grid control technologies.

EMF exposure from power lines can and should be reduced through undergrounding powerlines, improved transmission design, greater setbacks from schools, daycares and residential areas, and well established low EMF design. 

The Supraharmonic Monitoring Gap

Just as low-frequency noise (infrasound) from hyperscale data centers may not be audible but can still impact people living in surrounding communities, there are higher-frequency electrical harmonic distortions called supraharmonics that can impact power quality, without being measured or monitored. Utilities typically only monitor Total Harmonic Distortion (THD or VTHD) for low-order harmonics up to roughly 2–3 kHz (about the 40th–50th harmonic on a 60-Hz grid), but supraharmonics, generally defined as frequencies from about 2 to 150 kHz, are not routinely monitored. Despite this, they can be generated by electrical equipment in data centers as well as solar photovoltaic systems and other systems with high-frequency switching electronics. Front-end design measures to mitigate harmonics are important, however more needs to be done regarding oversight, measuring, monitoring, and research. 

Industry sources say that supraharmonics conducted emissions levels will only get worse due to the accelerated deployment of power electronics in or at the edge of the grid.

Power Quality Is Rarely Measured Where People Actually Experience It: Inside Homes

Harmonics can propagate through power lines and underground cables and into homes, schools, and workplaces. Bloomberg’s analysis highlighted a critical transparency gap: the widespread harmonic distortions identified across hundreds of thousands of homes were largely unknown to the public.

While engineering standards establish recommended limits for total harmonic distortion enforcement and reporting, requirements for reporting actual harmonic readings are limited, and most utilities only monitor power quality at substations rather than inside neighborhoods or homes. As a result, communities living near data centers have little information on the power quality and harmonic levels of the electricity delivered to their homes. 

What Can We Do? 

The Bloomberg investigation demonstrates that large-scale residential monitoring is technically feasible. The installation of monitoring networks in neighborhoods could provide continuous measurements of harmonic distortion, voltage fluctuations, and other indicators of grid stress in homes, schools, and workplaces. Such monitoring should also include extremely low frequency (ELF) magnetic fields, which can rise with heavy electrical loads and the expanded transmission infrastructure (powerlines and substations) associated with large data centers.

Residential measuring, monitoring, and improving transparency on power quality data should be a priority. Just as residents can easily check local air pollution levels through publicly available monitoring networks and online maps, similar monitoring systems could allow communities to view real-time power quality metrics, such as harmonic distortion and voltage fluctuations as well as magnetic field EMF levels, so that electrical grid impacts are openly monitored and addressed.

Resources

Articles on the Bloomberg Report

• AI data centers causing “distortions” in US power grid – Bloomberg – DCD
• The hidden disruptor of data center harmony: Vol 1 – DCD

Sheng, Y.; Zhang, C.; Zhu, Z.; Xu, H.; Wen, J.; Wang, R.; Yang, J.; Wang, Q.; Bu, S. Power for AI Data Centers: Energy Demand, Grid Impacts, Challenges and Perspectives. Energies 2026, 19, 722. https://doi.org/10.3390/en19030722

F. Portillo, M. Fernandez-Ros, N. Novas, R. M. Garcia, E. Viciana and J. A. Gazquez, “Monitoring Harmonic ELF Magnetic Fields in Data Centers with Sensor-Based Instrumentation,” in IEEE Transactions on Instrumentation and Measurement, doi: 10.1109/TIM.2026.3684676.

SupraHarmonics

Pinto J, Grasel B, Baptista J. Analysis of Supraharmonics Emission in Power Grids: A Case Study of Photovoltaic Inverters. Electronics. 2024; 13(24):4880. https://doi.org/10.3390/electronics13244880

Rajkumar, S., Balasubramanian, R. & Kathirvelu, P. A Comprehensive Review on Supraharmonics—The Next Big Power Quality Concern.Smart Grids and Energy 9, 15 (2024). https://doi.org/10.1007/s40866-024-00195-4

S. T. Y. Alfalahi et al., “Supraharmonics in Power Grid: Identification, Standards, and Measurement Techniques,” in IEEE Access, vol. 9, pp. 103677-103690, 2021, doi: 10.1109/ACCESS.2021.3099013 

J. Sutaria and S. K. Rönnberg, “Propagation of Supraharmonics in a Data Center With Different Operating Modes of UPS,” in IEEE Access, vol. 11, pp. 36823-36833, 2023, doi: 10.1109/ACCESS.2023.3266092

Sukič P, Dmitrašinović D, Štumberger G. Impact of Voltage Supraharmonics on Power Supply Units in Low-Voltage Grids. Electronics. 2025; 14(19):3918. https://doi.org/10.3390/electronics14193918

Powerside (company) Supraharmonics in the Grid: The Hidden Threat to Power Quality

EMF and Health 

Research on Harmonic Distortion and Health (also referred to as Dirty Electricity)

Milham S. Evidence that dirty electricity is causing the worldwide epidemics of obesity and diabetes. Electromagn Biol Med. 2014 Jan;33(1):75-8. doi:10.3109/15368378.2013.783853.  

Milham S, Stetzer D. Dirty electricity, chronic stress, neurotransmitters and disease. Electromagn Biol Med. 2013 Dec;32(4):500-7. doi: 10.3109/15368378.2012.743909.  

Milham, S. (2014). Hypothesis: the reversal of the relation between economic growth and health progress in Sweden in the nineteenth and twentieth centuries was caused by electrification. Electromagnetic Biology and Medicine, 33(1), 11–14. https://doi.org/10.3109/15368378.2013.783844

Havas, Magda. “Dirty electricity elevates blood sugar among electrically sensitive diabetics and may explain brittle diabetes.” Electromagnetic biology and medicine vol. 27,2 (2008): 135-46. doi:10.1080/15368370802072075

Havas M. Electromagnetic hypersensitivity: biological effects of dirty electricity with emphasis on diabetes and multiple sclerosis. Electromagn Biol Med. 2006;25(4):259-68. doi: 10.1080/15368370601044192. PMID: 17178585.