AI Data Centers Create Heat Islands, Warming Land Up to 16°F

New research reveals that AI data centers are generating "heat islands," causing surface temperatures to rise by an average of 3.6°F and up to 16.4°F in extreme cases, affecting areas within a 6.2-mile radius and over 340 million people globally.

Study Methodology

• Researchers from the University of Cambridge, led by Andrea Marinoni, analyzed 20 years of temperature data from remote sensors.
• The study mapped this data against over 6,000 large-scale AI "hyperscaler" data centers, mostly built in the last decade.
• To isolate heat impacts, data centers near dense urban areas were excluded, and seasonal trends, global warming, and other factors were filtered out.

Temperature Increases

• Average surface temperature increase near operational data centers: 3.6°F (2°C).
• Maximum recorded increase: 16.4°F (9°C) in extreme scenarios.
• Temperature rises extended up to 10 kilometers (6.2 miles) from data centers.

Affected Regions

• Bajio Region, Mexico: Unexplained temperature rise of approximately 3.6°F over 20 years, linked to data center growth.
• Aragon, Spain: Similar 3.6°F increase not replicated in neighboring provinces.
• Total impacted population: over 340 million people.

Expert Reactions

• Andrea Marinoni warned that planned data center expansion could have dramatic environmental, social, and economic impacts.
• Deborah Andrews, emeritus professor at London South Bank University, stated that the "rush for AI-gold" is overriding sustainable practices and systemic thinking.

Call for Action

• The study emphasizes the need for discussions on reducing AI's environmental footprint without hindering AI progress.
• Addressing heat emissions from proliferating data centers is critical amid global warming and extreme heat waves.