Data centers powering artificial intelligence are generating enough heat to measurably raise land surface temperatures, creating localized “heat islands” that already affect an estimated 340 million people. New research shows these facilities can increase temperatures by up to 9.1°C (16.4°F) in surrounding areas, a trend poised to worsen as AI demand surges.
The Scale of the Problem
The relentless expansion of AI is driving an unprecedented boom in data center construction. Real estate firm JLL projects that data center capacity will double between 2025 and 2030, with half of this growth fueled by AI. This exponential increase in energy consumption translates directly into increased heat output. Researchers at the University of Cambridge, led by Andrea Marinoni, quantified the impact by analyzing 20 years of satellite temperature data alongside the locations of over 8,400 AI data centers.
The study deliberately focused on data centers outside densely populated areas to isolate the thermal impact. The findings are stark: average land surface temperatures rose by 2°C (3.6°F) within months of a data center becoming operational, with extreme cases showing a 9.1°C (16.4°F) spike. Notably, the effect extends beyond immediate surroundings, with measurable temperature increases detected up to 10 kilometers away. Even at 7 kilometers distance, the intensity reduction was only 30%.
Why This Matters
The localized warming effect is not just an academic concern; it’s a potential public health and environmental issue. Over 340 million people live within 10 kilometers of these facilities, experiencing warmer temperatures than they would otherwise. The researchers found unexplained 2°C temperature increases in regions like Mexico’s Bajío region and Spain’s Aragon province over the past two decades, suggesting the impact is already being felt.
“This could become a huge problem,” says Marinoni, underscoring the urgency of addressing the thermal footprint of AI infrastructure.
Nuances and Future Considerations
While the study establishes a clear correlation between data centers and rising temperatures, some experts suggest further investigation is needed. Chris Preist at the University of Bristol notes the possibility that building materials and solar heating contribute to the effect, alongside computational heat. Regardless of the precise breakdown, the net result is undeniable: data centers are contributing to ground temperature increases.
The findings underscore the need for careful design and development of data center infrastructure. As AI continues to grow, ignoring this thermal impact will likely exacerbate localized climate change and potentially strain energy grids already under pressure.
