Artificial intelligence is rapidly transforming industries, economies and everyday life. But as governments and technology giants pour billions of dollars into building the infrastructure needed to power AI, researchers are warning of a little understood environmental consequence: data centres may be making the areas around them significantly hotter.
A new study titled The Data Heat Island Effect: Quantifying the Impact of AI Data Centers in a Warming World has found that AI data centres can increase local land surface temperatures by an average of more than 2°C after they begin operations. The researchers describe this phenomenon as the “Data Heat Island Effect”, drawing parallels with the urban heat island effect that causes cities to become warmer than surrounding rural areas.
The study analysed satellite temperature data collected between 2004 and 2024 across thousands of AI data centre locations worldwide. According to the researchers, the average increase in land surface temperature was 2.07°C, although some locations recorded temperature rises as high as 9°C.
The findings come at a time when the world is witnessing an unprecedented AI infrastructure boom. From the United States and Europe to India and the Middle East, hyperscale data centres are being built at record speed to meet the growing demand for AI applications. These facilities consume enormous amounts of electricity, much of which is ultimately released as waste heat into the surrounding environment.
What makes the study particularly significant is the scale of the impact. Researchers found that the warming effect is not limited to the data centre campus itself. Elevated temperatures can extend up to 10 kilometres from a facility, while a measurable increase of around 1°C can persist as far as 4.5 kilometres away.
The environmental implications could be substantial. Rising local temperatures can increase cooling requirements for homes and businesses, place additional pressure on electricity grids and worsen the effects of extreme heat events. Scientists have long linked heat islands to higher energy consumption, public health risks and stress on local ecosystems. The study suggests that AI driven infrastructure may now be emerging as another contributor to these challenges.
Perhaps most strikingly, the researchers estimate that more than 340 million people worldwide could be living in areas affected by data centre induced warming.
As the race to build AI infrastructure accelerates, the study argues that policymakers and industry leaders must look beyond carbon emissions alone. The environmental footprint of AI may not just be measured in megawatts and greenhouse gases, but also in the heat being added to local communities. In the age of artificial intelligence, the next climate challenge may be hiding in the very buildings powering the digital revolution.
