CAMBRIDGE, England (Michigan News Source) – As a Midwest heat wave approaches, the familiar chorus has already begun. Many Democrats and climate activists are once again pointing to rising temperatures as evidence of a climate crisis while pushing restrictions on everything from gas- powered leaf blowers and plastic straws to gas stoves, cow flatulence, and SUVs.

Yet one of the fastest-growing sources of concentrated energy use has largely escaped the same level of political scrutiny: artificial intelligence data centers. While critics have raised concerns about their enormous electricity demand, heavy water consumption, and round-the-clock noise, the significant amount of heat these massive facilities generate has received little to no attention.

Cambridge study finds AI data centers create “heat islands.”

MORE NEWS: Clinton Township Firefighter Among Three Killed Battling Colorado Wildfire

Now, new research suggests the environmental footprint of AI data centers may extend beyond electricity and water consumption. A new study by researchers led by the University of Cambridge found that AI data centers create what the authors call a “data heat island effect” by inducing local microclimate zones. Drawing on two decades of NASA satellite observations across the globe, the researchers found that after an AI data center begins operating, land surface temperatures in the surrounding area increase by an average of about 2 degrees Celsius (3.6 degrees Fahrenheit). The effect was measurable as far as 10 kilometers (6.2 miles) from the facility.

In some locations, land surface temperatures increased by as much as 9.1 degrees Celsius (16.4 degrees Fahrenheit). The researchers estimate that more than 340 million people worldwide live within 10 kilometers (6.2 miles) of the AI data centers included in the study, where they could be affected by the data heat island effect.

Not air temperature – but still heat.

The researchers measured land surface temperature – not the air temperature reported by weather stations. Their NASA satellite data captured the temperature of the land surface across 500-meter grid cells, which can include buildings, roads, parking lots, bare ground, and vegetation. Think of it like this: If you’ve ever stood behind a home’s outdoor AC unit or a restaurant exhaust fan, you’ve felt the hot air being expelled. A hyperscale AI data center does the same thing – but on a vastly larger scale, 24 hours a day.

Because of this, Cambridge concluded that hyperscale AI facilities are creating localized warming similar to the well-known urban heat island effect. In this scenario, cities are warmer than surrounding rural areas because concrete, asphalt, and buildings absorb and release heat during the day and slowly release it back into the environment, especially after sunset. That stored heat can contribute to warmer local air, higher air-conditioning costs, and added stress on vegetation by drying out soils more quickly. While the Cambridge study measured increases in land surface temperatures – not air temperatures – and did not examine those downstream effects, warmer surfaces are a well-known contributor to localized heat.

Michigan is going all-in.