In recent years, due to the rapid growth of artificial intelligence (AI) tools, companies around the world have built huge data centers. In addition to using excessive water and electricity, these data centers are creating “heat islands,” raising the temperature of the surrounding land by an average of two degrees Celsius, according to one study. This incident is further increasing the burden of heat on the lives of more than 34 crore people.
The study says: “We estimate that once AI data centers are operational, land surface temperatures will rise by an average of 2 degrees Celsius, creating smaller microclimate zones around them—an effect called the ‘data heat island effect’.” The study further states that in many cases, this increase in temperature can even reach 9 degrees Celsius. Andrea Marinoni, associate professor in the Earth Observation Group at Cambridge University and author of this study, said that her research was based on more than 6,000 data centers.
The increase in temperatures was consistent across the globe, and its impact was not limited to areas around data centers. Research showed that the increase in temperature caused by these AI data centers affected areas up to 9.9 kilometers (6.2 miles) away, affecting more than 340 million people.
AI water use
AI data centers rely on cooling systems that can use millions of liters of water—especially in hot regions. This raises many big questions regarding sustainability, especially in areas where there is already a problem of water scarcity.
According to scientists at the University of California, every 100 words of AI prompt requires approximately one bottle of water (or 519 ml). According to the Environmental and Energy Study Institute (EESI), a medium-sized data center can use approximately 110 million gallons of water per year for cooling. This is equivalent to the annual water use of approximately 1,000 households. Large data centers can “drink” up to 5 million gallons of water per day—about 1.8 billion gallons per year—equivalent to the water use of 10,000 to 50,000 people.