The rapid expansion of artificial intelligence technology has increased the demand for data centers, leading to concerns regarding their environmental and social effects. While AI energy efficiency is improving, the growth of energy-intensive applications like video generation has contributed to a 50% increase in electricity consumption by AI data centers last year, according to the International Energy Agency. Public sentiment reflects these concerns, with a September CNN poll finding nearly 90% of Americans are more concerned than excited about the rise of these facilities.
Data centers currently account for approximately 0.5% of global electricity consumption, a figure Our World in Data projects will rise to 1.5% by 2030. In the United States, where nearly half of the world's data center power is used, Lawrence Berkeley National Laboratory estimates these facilities could consume up to 12% of the nation's electricity by 2028. To meet this demand, companies are increasingly developing private energy infrastructure, such as natural gas turbines, to avoid straining local grids and increasing consumer utility bills.
Environmental impacts also extend to water usage, though estimates vary. OpenAI CEO Sam Altman stated a ChatGPT query uses about one-fifteenth of a teaspoon of water, while other researchers estimate a 100-word AI email requires roughly two cups. Beyond resources, residents in areas like Vineland, New Jersey, have filed lawsuits over noise pollution from cooling systems and generators. Additionally, an Arizona State University study found that heat exhaust from data centers in Phoenix warmed downwind neighborhoods by as much as 4 degrees Fahrenheit.
For local residents, the day-to-day changes include increased noise from cooling equipment and potential "heat island" effects. In terms of health, a recent study suggested the public health burden of US data centers, caused by air pollutants like nitrogen oxides and fine particulate matter from on-site power generation, could reach $20 billion by 2028. These facilities also drive demand for critical minerals, placing the AI industry in competition with the wind and solar sectors for raw materials needed for semiconductors and batteries.
Next steps for the industry involve balancing technological demand with resource management. Experts like Kaveh Madani of the United Nations University state that the ultimate impact will depend on regulatory choices regarding facility location and energy transparency. While some companies are developing new cooling technologies to reduce water consumption, many projects remain in development. In Imperial, California, residents held protests on July 18 against a proposed 950,000 square foot data center, citing health and environmental concerns. No specific federal deadlines for new data center regulations were reported.