Major tech companies like Google, Amazon, and Microsoft are building massive AI data centers, known as hyperscalers, to support artificial intelligence. However, these facilities are significantly raising the temperature in their surrounding areas, according to a study led by Cambridge University. The phenomenon, termed 'Data Heat Island Effect,' shows an average temperature increase of 2 degrees Celsius, with some areas experiencing up to 9 degrees Celsius rise.
These data centers consume vast amounts of electricity, with the International Energy Agency reporting that global data centers used 415 terawatt-hours in 2024, equivalent to 1.5% of global electricity supply. This demand is expected to double by 2030. Hyperscalers, which require 100 to 300 megawatts of power per second, also demand significant water resources for cooling, with one 100-megawatt center needing 250 million liters annually.
As of June 2026, there are over 11,600 data centers worldwide, with the United States hosting the most at over 4,300. The UK, Germany, and France lead in Europe, while China and India top the list in Asia. The number of hyperscalers has grown from 700 in 2021 to 1,297 currently. Researchers warn that the heat from these centers can affect local populations and ecosystems, emphasizing the need for environmentally friendly AI development.
