Bloomberg Investigation Shows AI Data Centers Warming Nearby Neighborhoods
Researchers found air-cooled facilities can raise nearby air temperatures by up to 4 degrees Fahrenheit and increase cooling demand in surrounding neighborhoods.
- Arizona State University researchers found air-cooled data centers near Phoenix create thermal plumes that raise neighborhood temperatures by up to 4 degrees Fahrenheit, revealing an emerging urban thermal hazard.
- Driven by the generative AI boom, large data centers consume massive amounts of water and electricity, with a single facility using 5 million gallons of water daily and straining electrical grids.
- A 30 megawatt data center emits energy comparable to 25,000 or 35,000 homes, David Sailor, director of ASU's School of Geographical Sciences and Urban Planning, noted, intensifying air conditioning demand and creating a "feedback loop."
- Public opposition is mounting, with seven in 10 Americans opposing local AI data centers; more than 140 local groups have blocked or delayed projects worth more than $60 billion over roughly a year.
- Yuan Yao, the Manufacturer's Association Professor of Industrial Ecology and Sustainable Systems, suggests that transparent reporting on life cycle environmental impacts is critical for reconciling infrastructure growth with sustainability.
10 Articles
10 Articles
New Evidence Data Centers May Cause Hotter Weather
Data centers continue to spread across the U.S. landscape as demand for artificial intelligence, social media and digital services surges. Their impact on energy and water supply is well documented, but the first in a series of new studies has revealed another potential impact of data centers: They may be creating enough heat to affect temperatures around them and produce data center heat islands, or DCHIs.
AI May Be Ruined Data Centers
Wall Street is getting cold feet on data centers, and communities across America are pushing back hard against new construction. The technology powering the AI revolution may be colliding with limits no algorithm can optimize away.
Coverage Details
Bias Distribution
- 86% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium














