UN report: AI could double electricity use by 2030, straining grids and water supplies

A new United Nations report warns that artificial intelligence could double its electricity consumption by 2030, reaching about 3% of global power use. That would produce greenhouse gas emissions equivalent to the entire United Kingdom. The report also flags a hidden cost: AI data centers could require 9.3 trillion liters of water for cooling, more than the annual drinking water needs of the global population. The report points to the Jevons paradox, where efficiency gains in AI models will likely lead to more total usage, not less. Data centers already consumed as much electricity as Saudi Arabia last year. If that doubles by 2030, offsetting the carbon would need 6.7 billion trees grown over a decade. The report calls for mandatory environmental disclosures at the model and task level, plus full value-chain governance from mineral sourcing to recycling. Only 32 nations host AI-specific cloud infrastructure, with 90% of capacity in the US and China. That means countries consuming AI services will bear a disproportionate environmental burden from mineral extraction and e-waste. For anyone tracking grid decarbonization or climate finance, this report makes clear that AI growth is becoming a major energy and water story that regulators and investors cannot ignore.