A new major study has identified a phenomenon called the “Data Heat Island Effect,” highlighting how artificial intelligence data centers may contribute to global warming and localized climate change.
The study, published in 2026, was conducted by international researchers from institutions including University of Cambridge, Nanyang Technological University, and City University of Hong Kong, among others. In the study, they analyzed 20 years of NASA satellite data from 2004 to 2024 and cross-referenced more than 8,400 AI data centers worldwide.
Researchers defined the “Data Heat Island Effect” as localized climate zones created by data centers that may intensify global warming and climate change.
The findings showed that AI-focused “hyperscalers” — massive facilities powering modern artificial intelligence — create localized heat islands, with land surface temperatures increasing by an average of 2°C after the start of AI data center operations.
The study also documented extreme heat spikes in some locations, with temperatures rising as high as 9.1°C.
Researchers found that the geographic impact of AI data centers extends beyond the facilities themselves, reaching up to 10 kilometers away, with the most intense heat concentrated within a seven-kilometer radius.
According to the study, around 340 million people worldwide live in areas near these facilities.
Unlike traditional data centers, AI facilities require thousands of high-performance GPUs, or graphics processing units, operating at full capacity for extended periods. This generates large amounts of thermal energy that cooling systems must expel from buildings into surrounding communities.
The study highlighted two regions where the “Data Heat Island Effect” was particularly evident and could not be fully explained by global warming alone: Bajío, Mexico, and Aragón, Spain.
Researchers said temperatures in the rapidly expanding data center hub of Bajío rose by about 2°C over two decades, directly correlating with the construction of new facilities.
A similar pattern of localized warming was observed in Aragón, where neighboring provinces without data centers did not experience the same temperature increases.
“Our results show that the data heat island effect could have a remarkable influence on communities and regional welfare in the future, hence becoming part of the conversation around environmentally sustainable AI worldwide,” the researchers said.
The study warned that the phenomenon could have implications for public welfare, healthcare, energy demand, and population resilience.
“Although the impact of data heat islands can be intense, advances in semiconductor technology, energy materials, computer science, and electrical engineering can help mitigate their effects,” the researchers added.
Researchers also warned that as data center energy consumption is expected to rise sharply in the coming years, the “Data Heat Island Effect” could become an increasingly significant factor in environmental and industrial sustainability worldwide.
