The rapid expansion of Big Tech’s data centers has led to significant public health costs in the United States, with a recent study estimating these costs to have reached $5.4 billion over the past five years. The research, conducted by UC Riverside and Caltech, highlights the growing health impacts of the energy-intensive infrastructure that powers artificial intelligence (AI) and other tech services.
According to the findings, air pollution from data centers—caused by the immense energy demands required to operate them—has been linked to a rise in illnesses such as cancer, asthma, and other respiratory issues. The researchers found that the cost of treating illnesses related to this pollution amounted to $1.5 billion in 2023 alone, marking a 20% increase from the previous year. The total cost since 2019 stands at an alarming $5.4 billion.
The Financial Times notes that this trend is expected to worsen as demand for AI services continues to soar. The surge in AI development requires massive computing resources, which directly translates to higher energy consumption in data centers. Industry giants like Microsoft, Alphabet, Amazon, and Meta have already forecast AI-related spending to exceed $320 billion this year, up from $151 billion in 2023. Moreover, OpenAI and SoftBank recently unveiled plans for a $500 billion joint venture to build AI infrastructure in the U.S., further exacerbating the environmental and health costs associated with the sector’s expansion.
The research from UC Riverside and Caltech utilized a widely recognized modeling tool from the U.S. Environmental Protection Agency (EPA), which translates air quality and human health impacts into a monetary value. The results suggest that Google generated the highest health costs—$2.6 billion—between 2019 and 2023, followed by Microsoft at $1.6 billion and Meta at $1.2 billion. Other major tech firms, like Amazon, were not included in the study due to the lack of publicly available data on their energy consumption.
The pollution from data centers primarily stems from their reliance on electricity, much of which is still sourced from fossil fuels. Additionally, back-up generators powered by diesel fuel contribute to air pollution, while discarded hardware like chips can release harmful chemicals into the environment. This research focused on pollution generated in the specific areas where the data is processed, an approach known as “location-based” accounting, rather than accounting for market-based instruments like renewable energy certificates, which tech companies purchase to offset their emissions.
While companies like Google, Microsoft, and Meta have responded by emphasizing their clean energy purchases and sustainability efforts, critics argue that the health cost estimates remain valid. Google, in particular, claimed the research overstated its health costs and did not account for its clean energy purchases, which it says provide around 64% of its energy needs carbon-free. However, researchers counter that these offset measures do not change the localized impact of air pollution.
The study also revealed that lower-income households in areas like West Virginia and Ohio have disproportionately suffered from the health impacts of data center pollution, due to the locations of these facilities. Shaolei Ren, an associate professor at UC Riverside, suggested that tech companies could alleviate this public health threat by strategically siting data centers in less populated regions.
A separate report by Berkeley Lab, supported by the U.S. Department of Energy, projected that energy use by U.S. data centers, which accounted for about 4% of the nation’s electricity consumption in 2023, will rise significantly in the coming years. By 2028, it is expected to reach between 7% and 12%, largely driven by the growing demand for AI services.
Environmental experts like Antonis Myridakis from Brunel University London emphasize the importance of addressing the pollution caused by AI’s energy demands. “It is an important factor contributing to air quality and public health, and it is not something we can ignore,” he said.
The findings underscore a critical dilemma facing Big Tech: balancing the rapid growth of AI infrastructure with the health and environmental costs that come with it. The Financial Times highlights that, without significant action, the industry’s expansion could have long-lasting effects on both public health and the environment.
