Jakarta, INTI - The backbone of artificial intelligence (AI), massive data centers, undeniably consumes enormous amounts of energy. Beyond energy use, these facilities can also create heat islands due to their operations.
A study titled “The Data Heat Island Effect: Quantifying the Impact of AI Data Centers in a Warming World” reveals that AI-based data centers can raise surrounding land temperatures by up to 16 degrees Fahrenheit (0.5°C).
According to the research, surface temperatures typically increase by 3.6 degrees Fahrenheit once a data center becomes operational, and in extreme cases, temperatures can rise as much as 16.4 degrees Fahrenheit.
In Baijo, Mexico, a hub for data center construction, researchers observed an unexplained temperature increase of 3.6 degrees Fahrenheit over the past 20 years. Similar patterns were recorded in Aragon, Spain, home to Europe’s largest AI data centers, with temperature rises not seen in neighboring provinces.
This surface temperature increase is not isolated to these locations, it has been observed globally. The impact extends beyond the immediate facility, affecting areas up to 6.2 miles (9.98 km) away and potentially impacting 340 million people.
Researchers analyzed 20 years of temperature data using remote sensors, mapping them against hyperscaler AI data centers, facilities that can exceed one million square feet and house thousands of servers. The study focused on 6,000 data centers located in remote areas to isolate temperature changes from other human activities, such as manufacturing or residential energy use, while controlling for seasonal and global warming trends.
Data Center Expansion and Environmental Concerns
Andrea Marinoni, one of the researchers, warned that the projected expansion of data centers could have dramatic consequences for society, including environmental, economic, and social impacts. He emphasized exploring alternative approaches to reduce AI’s environmental footprint without compromising its benefits to human progress.
Deborah Andrews, Professor Emeritus of Sustainable Design and Circular Economy at London South Bank University, noted that AI development is progressing far faster than broader sustainable systems and best practices.
Meanwhile, Ralph Hintemann, Senior Researcher at the Borderstep Institute for Innovation and Sustainability, highlighted that the largest climate impact of data centers stems from the emissions of their power plants. “When it comes to climate change, the emissions generated to power data centers remain the most concerning aspect,” he said.
Conclusion
AI data centers are driving unprecedented energy consumption and creating localized heat islands, affecting both the environment and nearby populations. While operational heat is a concern, the most significant climate impact comes from the emissions of the power plants supplying these centers. Balancing AI growth with sustainable energy and environmental practices remains a critical challenge.
Read more: Meta Funds Construction of 7 Gas-Fired Power Plants to Meet Data Center Energy Needs