Warming up the pace of AI took on a much clearer meaning.
AI devices are heating up the surrounding areas, creating a “data heat island effect” that has the potential to affect hundreds of millions of people living nearby, a new working paper has found.
Using a land surface temperature dataset produced by NASA, a research team led by the Department of Computer Science and Technology at the University of Cambridge found that from 2004 to 2024, locations near more than 6,000 data centers around the world saw an average global warming of about 2 degrees Celsius, or 3.6 degrees Fahrenheit. In some cases, nearby temperatures increased by 9 degrees Celsius, or 16.4 degrees Fahrenheit. The researchers estimated that these heat islands could be felt 6.2 kilometers away from utilities, affecting about 343 million people worldwide.
“The effect of data heat islands can have a dramatic impact on society and the well-being of regions in the future,” said the study, which has not yet been peer-reviewed.
Data centers, which store and process large amounts of data to train AI, have become the basis of AI-related spending, and capital expenditures for these services are predicted to reach $760 billion by 2026, according to BloombergNEF estimates, up from $450 billion last year. Hyperscalers like Alphabet doubled their spending on data centers this year, and parent company Google plans to invest $185 billion in AI infrastructure. The capital expenditure for these large technology firms exceeds the GDP of entire countries such as Sweden.
The effects of weather on the information center
The power required to run these data centers is huge. Modern AI runs on clusters of tens of thousands of graphics processing units (GPUs), which generate massive amounts of heat that require air and water to cool. With some modern data centers spanning hundreds or thousands of acres, the energy needed to power and cool the buildings can exceed a gigawatt, roughly enough to power between 750,000 and 1 million homes.
The amount of energy required has raised concerns about the environmental impact of data centers, as well as the disruptions they may cause to nearby residents. These facilities can cause noise pollution, producing noise levels above 90 decibels. Loud volumes above 85 decibels are considered harmful to hearing. In arid regions, the large use of water for cooling data centers has raised concerns about the possibility of drought.
Increased energy consumption has also taxed the US’s aging grid system – and combined with extreme weather and rising natural gas costs, electricity bills will rise by 7% from December 2025, according to Goldman Sachs analysts. These increased energy costs will be passed on to consumers, especially low-income Americans, because when businesses like restaurants struggle with increased energy costs, they may increase prices, including food, to offset those higher costs.
“Income and spending will be greater for low-income households because electricity accounts for a large portion of their consumption, as well as for households in areas with high data services where the regional energy market will tighten,” Goldman Sachs economists Manuel Abecasis and Hongcen Wei wrote in a note to clients in February.
Indeed, the heat island study has raised criticism about how much energy use data centers actually have on the environment. Some independent researchers have noted that most of the increased ground heat from the building of the data center comes from the energy required to build any building where the bare soil and plants once lived, not the heat created by the work of the data center.
Risks of general access to data centers
These implications are significant, experts say, especially given the question marks over the sustainability of AI spending. According to the latest Moody’s analysis, the total commitment of spending made by Alphabet, Amazon, Meta, Microsoft, and Oracle, about two-thirds of it, $ 662 billion, is planned for fees related to the data center that is about to start. These hyperscalers issued $121 billion in new debt in bonds last year alone.
The risks of expanding the data network have been increased by the ongoing war in Iran. Not only has the country threatened Nvidia, Apple, Microsoft and Google with data center attacks, but the reduced energy business has disrupted data service supply chains.
As the expansion of AI infrastructure grows, so do the financial and environmental risks associated with it.
“AI infrastructure is a challenge of energy and healing embedded in the opportunity of the digital economy,” Lee Poh Seng, a professor in charge of thermal systems at the National University of Singapore (NUS), said earlier. Good luck.
Researchers, however, see a way forward to reduce the heat island effect of AI buildings. They suggest software-based solutions that increase the efficiency of computing systems and therefore require less power. Hardware-based solutions include improvements to integrated circuitry, or the structure of the chips themselves, to help recover energy, as well as implementing hybrid cooling systems that combine “liquid cooling at the chip location and air cooling across the board.”
“Although the impact of data heat islands can be strong (as mentioned earlier),” a recent Cambridge study said, “technological progress in the semiconductor industry and energy equipment, as well as advances in computer science and electrical engineering, can be used to reduce their effects.”
#Data #centers #hot #heat #island #effect #raises #temperature #miles #Good #luck