AI Data Centers: How Much Heat Do They Generate and Where Are They Located?
Marium Ali, Hanna Duggal
A new study reveals that AI data centers raise surrounding land temperatures by an average of 2 degrees Celsius, with some areas seeing increases up to 9 degrees Celsius. The research, led by the University of Cambridge, warns that over 340 million people live within 10 kilometers of a data center and may be affected by this 'data heat island effect.'
Tech giants are racing to build infrastructure for artificial intelligence (AI). But mounting evidence suggests that AI hyperscalers—large-scale cloud providers such as Google, Amazon, and Microsoft—are also heating the ground around them.
A study led by researchers at the University of Cambridge found that the land surface temperature around AI data centers rises by an average of 2 degrees Celsius (3.6 degrees Fahrenheit), with some areas recording increases of up to 9 degrees Celsius (16.2 degrees Fahrenheit). The researchers call this the 'data heat island effect.'
Energy consumption of AI data centers
Every time someone uses ChatGPT, Gemini, or Claude, the request is processed at a data center—a vast facility filled with specialized computers running 24/7. AI data centers use powerful chips that perform thousands of parallel calculations, and continuously running large models makes them far more energy-hungry than standard web-serving servers.
According to the International Energy Agency (IEA), data centers consumed about 415 terawatt-hours (TWh) of electricity in 2024, roughly 1.5% of global supply, with growth of about 15% annually over the past five years. That figure is forecast to nearly double, reaching 945 TWh by 2030.
Among them, hyperscale data centers are the most energy-intensive. According to IBM, they typically have at least 5,000 servers and occupy a minimum of 10,000 square feet (930 square meters). These facilities often require 100 to 300 megawatts of electricity to operate—enough to power hundreds of thousands of homes. That energy generates enormous heat, which must be managed through advanced liquid-cooling systems that consume large amounts of water.
A report by the UK government's digital sustainability advisory body found that a 100-megawatt hyperscale data center can consume about 2.5 billion liters of water per year—equivalent to the annual needs of 80,000 people.
The data center construction boom
As of June 2026, there were more than 11,600 data centers operating worldwide. The majority are concentrated in the United States, with over 4,300 centers. Europe is the second-largest hub, led by the United Kingdom (more than 540), Germany (more than 520), and France (more than 390). In Asia, China (more than 360) and India (more than 300) lead the region, while Southeast Asia is expanding rapidly. According to Synergy Research Group, the number of hyperscale data centers worldwide has nearly doubled since 2021, from 700 to 1,297.
Heat generated by AI data centers
The study from Cambridge, Nanyang Technological University, and others shows that land surface temperatures around AI data centers rise by an average of 2 degrees Celsius after they become operational, with effects detectable up to 10 kilometers (6 miles) away. Using NASA satellite data, the researchers measured global land surface temperatures from 2004 to 2024 and cross-referenced them with more than 11,000 AI data center locations. The study focused on 6,733 centers located outside densely populated areas and found temperature increases ranging from 0.3 degrees Celsius to 9.1 degrees Celsius.
The research indicates that more than 340 million people living within 10 kilometers of a data center could be affected by the temperature rise—an impact the researchers describe as having 'significant implications for community and regional well-being' and one that should be included in the global conversation about environmentally sustainable AI.
Major investment projects
Global investment bank Goldman Sachs expects total capital spending from 2025 to 2030 by the four largest hyperscalers (Microsoft, Amazon, Alphabet, and Meta) to reach $5.3 trillion. Major upcoming projects include: Meta's $27 billion Hyperion campus in Louisiana; Microsoft's multi-phase, $20 billion data center expansion project in Wisconsin; Amazon's $25 billion investment in data center infrastructure in Mississippi; Google's $15 billion Project Spade in Missouri; and Oracle's Project Stargate in Texas, an AI supercluster dedicated to OpenAI with total capacity between 1.2 GW and 2 GW.
