AI Could Soon Use 3% of Global Electricity and More Water Than the World Drinks, Warns UN

HAMILTON, June 9: Artificial intelligence could account for nearly 3 per cent of the world’s electricity consumption by 2030 and use more water for cooling than the annual drinking water needs of the global population, according to a new United Nations report that highlights the growing environmental costs of AI.

The report cautions that improvements in AI efficiency may not necessarily reduce overall resource consumption. Instead, it says lower costs and greater accessibility could drive wider adoption, increasing total energy and water use despite technological advances.

Drawing on the economic principle known as the “Jevons paradox,” the report argues that gains in efficiency often lead to higher overall consumption. Named after 19th-century economist William Stanley Jevons, the theory suggests that making a resource more efficient to use can increase demand for it rather than reduce it.

According to the report, AI’s energy consumption could double by the end of the decade, resulting in carbon emissions comparable to those of the United Kingdom. Data centres supporting AI systems could require an estimated 9.3 trillion litres of water annually and occupy land nearly ten times the size of Mexico City.

The report notes that data centres already consumed as much electricity in 2024 as Saudi Arabia, the world’s 11th-largest electricity user. If projected energy demand materialises, offsetting the associated carbon emissions would require the equivalent of growing 6.7 billion trees over a decade.

Beyond environmental concerns, the report highlights growing inequalities in the global AI ecosystem. It states that only 32 countries host AI-specific cloud infrastructure, with around 90 per cent of that capacity concentrated in the United States and China.

The report warns that developing countries often bear the environmental consequences of mineral extraction and electronic waste while lacking control over the technologies driving AI growth. This, it says, could widen the digital divide between nations that develop AI systems and those that primarily consume them.

To address these challenges, the UN report calls for responsible AI governance across the entire technology lifecycle, from mineral sourcing and manufacturing to recycling and disposal. It recommends greater transparency regarding AI’s environmental impact, routine disclosure of energy use and emissions, and integration of AI demand into climate and energy planning.

The report also urges governments to balance technological innovation with environmental stewardship, warning that a light-touch regulatory approach may fail to address the sector’s rapidly growing resource demands.

With countries increasingly adopting AI across public services and government functions, the report stresses that sustainability considerations must become central to future AI development and deployment strategies. (Agencies)