The rapid expansion of data centers is triggering a growing public health crisis, according to a new analysis by Harvard researchers.
The explosive growth of artificial intelligence has created massive electricity demand, leading many data centers in states such as Virginia, Texas, and Mississippi to install on-site fossil fuel-powered backup generators. While the tech industry emphasizes grid reliability, public health experts are raising alarms about the local health impacts of these facilities.
A recent study led by Harvard biostatistician Michael Cork found that emissions from just one data center in Loudoun County, Virginia, could generate up to $99 million in annual health damages, including premature deaths, respiratory illnesses, and increased hospitalizations. These large fossil-fuel plants are often built near residential neighborhoods, frequently in socioeconomically vulnerable areas, exposing nearby communities to elevated levels of fine particulate matter and smog.
In addition to air pollution, residents are facing another serious issue: constant low-frequency noise from massive cooling systems, fans, and generators. People living near data centers in the US and Europe report chronic sleep disruption, headaches, anxiety, and other stress-related symptoms.
Medical research has long linked prolonged exposure to environmental noise with higher risks of cardiovascular disease, elevated stress hormones, and impaired cognitive development in children.
As tech companies continue to rapidly scale their infrastructure, critics argue that the significant but often overlooked public health and environmental costs of AI are being sidelined in regulatory and permitting decisions.
[Feldscher, K. (2026). Analyzing air pollution health, economic risks from AI data centers. Harvard T.H. Chan School of Public Health]