A recent study conducted by researchers from UC Riverside and Caltech has revealed that air pollution caused by data centers operated by Google, Microsoft, and Meta resulted in an estimated $5.4 billion in public health costs in the United States between 2019 and 2023. This pollution, primarily stemming from the high energy demands of these facilities, has been linked to illnesses such as cancer, asthma, and other respiratory issues.
Among the three companies, Google was responsible for the highest share of health-related costs at $2.6 billion, followed by Microsoft at $1.6 billion and Meta at $1.2 billion. The energy consumption of these data centers is largely supplied by electricity generated from fossil fuels, which emit fine particulate matter and nitrogen oxides—pollutants known to cause significant health problems. Additionally, backup diesel generators used during power outages contribute further to air pollution.
The study highlights that low-income communities located near data centers or power plants have been disproportionately affected by this pollution. For instance, emissions from data centers in Northern Virginia have drifted into neighboring states like Maryland and New Jersey, creating regional public health costs of up to $260 million annually. Despite tech companies’ claims of sustainability efforts—such as purchasing renewable energy certificates—these measures do not mitigate the localized air quality impacts caused by their operations.
The findings underscore the urgent need for stricter regulations on data center emissions and strategic siting of facilities to minimize health impacts. Researchers recommend that tech companies adopt transparent reporting standards for air pollution and provide compensation to affected communities. As demand for AI services continues to grow, the associated environmental and public health challenges are expected to escalate unless significant changes are made.