The rapid expansion of artificial intelligence technology is driving an environmental crisis that extends far beyond carbon emissions, according to a study by the United Nations.
The study, published by researchers from the United Nations University (UNU), warns that the massive data centers required to power AI are placing unprecedented, interconnected pressures on global electricity, water, and land resources.
Global data center electricity consumption could reach up to 945 terawatt-hours by 2030. This projected demand is nearly three times the current combined power usage of Pakistan, Bangladesh, and Nigeria—nations with a collective population exceeding 650 million people, the UN study said.
However, researchers cautioned that electricity consumption represents only a fraction of the total environmental footprint. Because power generation and data center cooling infrastructure are highly resource-intensive, the rising demand for electricity directly accelerates water and land scarcity.
The study projects that by 2030, the annual water required to sustain AI data centers will match the basic liquid sustenance needs of 1.3 billion people.
Furthermore, the physical footprint of these facilities is expected to exceed 14,500 square kilometers, an area roughly twice the size of the metropolitan area of Jakarta, Indonesia.
The study criticizes current environmental assessments of AI for focusing too narrowly on carbon footprints, as it also leaves “water footprint” and “land footprint” as each electricity unit is also associated with land and water.
Researchers found that some initiatives designed to mitigate carbon emissions—such as transitioning to specific renewable energy sources—can inadvertently increase regional water consumption and land degradation.
“Solutions seen as “green” in one sense may worsen pressures in others, particularly in regions already facing resource scarcity. For example, switching to certain renewable energy sources may reduce carbon emissions but can significantly increase water consumption and land use,” the report said.
The report also said that daily AI operational usage, aside from the training of AI models, accounts for 80% to 90% of AI’s total energy demand. With users worldwide now generating an estimated 2.5 billion prompts per day, operational demand consumes hundreds of gigawatt-hours of electricity annually.
“Generating a single AI image can require more than a thousand times the energy of simple text classification, while video generation demands even greater resources,” the UN said.
The UN has urged governments to integrate AI infrastructure into energy, water, and land-use planning.
It also encouraged companies to design systems that reduce resource consumption, while users are advised to choose “lower-impact” applications where possible.
The UN clarified that it is not opposed to AI or technological innovation, but is calling for urgent action to address its growing environmental impact worldwide.
