AI is now a major part of daily life, and many companies are racing to build it into their products and services. But that demand needs a huge amount of computing power, and researchers say the environmental cost may be larger than many people realize.
The wider debate about climate change remains tense, especially as the Trump administration has taken a different approach to the environment than several past governments. Against that backdrop, the effect of AI infrastructure is becoming another issue for scientists, officials, and nearby residents to watch.
The findings came from researchers at the University of Cambridge. The impact could affect more than 340 million people around the world, based on where these centers are being built and how heat spreads around them.
Andrea Marinoni, an associate professor at the British university, told CNN that much more work is needed to understand the big gaps in what experts know about how data centers affect the planet.
The University of Cambridge team analyzed temperature data over a 20-year period in places where AI data centers have appeared in recent years.
To keep the results clearer, the scientists looked at about 6,000 data centers located away from dense city areas. That mattered because surface temperatures in crowded urban zones can already be affected by factories, traffic, buildings, and other sources of heat.
On average, they found that surface temperatures rose by 3.6 degrees Fahrenheit after the centers opened. In some cases, the increase was far higher, with a rise of 16.4 degrees Fahrenheit recorded.
Deborah Andrews, emeritus professor of design for sustainability at London South Bank University, spoke to CNN after the striking findings were shared publicly.
Andrews added to the outlet: “The ‘rush for AI-gold’ appears to be overriding good practice and systemic thinking and is developing far more rapidly than any broader, more sustainable systems.”
Ralph Hintemann, another expert in the field, also questioned parts of the findings. He described the higher temperature figures as interesting, but also very high, suggesting that more study will be needed before the full picture is clear.
That kind of software could, in theory, shift tasks toward cleaner energy sources or times when the grid is under less pressure. It would not solve the whole problem, but it could be one step toward making AI systems less wasteful.
The study adds to a growing concern around the hidden cost of AI. While the tools may feel digital and distant to users, the data centers behind them take up real space, use real power, and may be changing temperatures for people who live nearby.
