According to the latest research released by the University of Waterloo in Canada on November 17, 2025, it is generally believed that artificial intelligence (AI) will bring a heavy burden to the climate, but in fact, current AI applications only have a minimal impact on global greenhouse gas emissions, and even help promote environmental protection technology and economic growth.
The research was jointly conducted by scientists from the University of Waterloo and the Georgia Institute of Technology. Researchers combined data from the U.S. economic system and estimates of AI penetration in various industries to predict the possible environmental impact of future AI expansion.
According to the U.S. Energy Information Administration, 83% of U.S. energy still relies on high-emission fossil fuels such as oil, coal, and natural gas. This study pointed out that although the United States’ AI-related electricity consumption is the same as Iceland’s national energy consumption, the overall scale is still not enough to have a significant impact at the national or global level.
Professor Juan Moreno-Cruz, School of the Environment at the University of Waterloo and Canada Research Chair in Energy Transition, said: "The incremental energy consumption is not evenly distributed, but has a greater impact in specific areas that supply power to data centers. The power generation and emissions in some areas may even double, but from a larger perspective, AI energy consumption is not outstanding overall."
Although the paper did not examine the impact of data centers on the local economy in detail, the results are still optimistic. Moreno-Cruz pointed out: "For those who believe that AI will become a major climate issue and argue that it should be avoided, we provide a different perspective. AI will have a limited impact on the climate and can be used to develop green technologies and improve existing solutions."
In order to draw conclusions, the research team conducted an analysis from the perspectives of various economic industries, jobs, and the proportion of possible replacements by AI. Moreno-Cruz and his collaborator Anthony Harding plan to promote this research method to other countries in the future to further evaluate the impact of AI applications on a global scale.
Compiled from /ScitechDaily