A new study conducted by the UN think tank, the United Nations University (UNU), has warned that the rapid expansion of artificial intelligence could place unprecedented pressure on global energy, water, and land resources, with impacts expected to be particularly significant for developing regions in Asia.

The report titled The Environmental Cost of Artificial Intelligence: Carbon, Water and Land Footprints, published by the UNU Institute for Water, Environment and Health, found that AI’s environmental footprint extends far beyond carbon emissions and is growing at a pace that could strain natural resources worldwide.

According to the study, AI-related electricity consumption is projected to increase sharply by 2030, requiring nearly three times the annual power consumption of Pakistan, Bangladesh, and Nigeria combined — countries with a total population of about 650 million people.

The report estimates that AI systems could generate up to 400m tonnes of carbon dioxide emissions annually by the end of the decade, roughly equivalent to the UK’s current emissions.

The study also highlights water consumption as a major concern. By 2030, AI infrastructure could require around 9.3 trillion litres of water annually, an amount comparable to the domestic water needs of 1.3 billion people in sub-Saharan Africa.

Land occupied by AI-related infrastructure, including data centres and energy facilities, could exceed 14,500 square kilometres.

UNU researchers warned that while the benefits of AI are concentrated in a handful of countries, the environmental costs are likely to be borne disproportionately by other regions, including parts of Asia where water stress and energy demand are already rising.

The report calls for greater transparency from technology companies, stronger environmental governance, and international cooperation to ensure AI development remains sustainable.