A new report from the United Nations has raised serious alarms about the environmental impact of artificial intelligence (AI), warning that by 2030, AI could consume up to 3% of the world's electricity and require more water than the global population needs for drinking. The report, released on June 4, 2026, challenges the common argument that efficiency gains in AI will reduce its resource demands, instead pointing to a phenomenon known as the Jevons paradox.
The Jevons Paradox and AI
The Jevons paradox, named after 19th-century economist William Stanley Jevons, describes a situation where technological improvements that increase the efficiency of a resource's use lead to a rise, rather than a fall, in total consumption. Jevons observed this with coal in England: as steam engines became more efficient, coal use increased because the lower costs expanded its applications. The UN report predicts a similar effect for AI. As AI models become more efficient and cheaper, they will likely be used in more applications and at higher volumes, potentially erasing any environmental savings from efficiency improvements.
Scale of the Problem
According to the report, data centres already consumed as much electricity in 2025 as Saudi Arabia, the world's 11th largest electricity consumer. If energy use doubles by 2030 as projected, the carbon footprint would require 6.7 billion trees grown over a decade to offset. Additionally, data centres would need 9.3 trillion litres of water for cooling, surpassing the annual drinking water needs of the global population. Land use would also surge, requiring an area nearly ten times the size of Mexico City.
Structural Inequity and the Digital Divide
The report highlights that only 32 nations host AI-specific cloud infrastructure, with 90% of that capacity in the United States and China. This concentration deepens the digital divide, as countries that consume AI often bear the environmental costs of mineral extraction and e-waste without reaping the full benefits. The UN calls for global cooperation to address these inequities.
Pathways to Responsible AI
The report outlines a roadmap for responsible AI use based on transparency, efficiency by design, equity, lifecycle responsibility, and sustainable use. It stresses that the environmental impact of AI depends on both how much it is used and how it is used, as different tasks and models have varying energy and resource costs. The UN recommends full value-chain governance, from mineral sourcing to recycling, and urges governments to incorporate AI's projected demand into climate and energy planning.
National Examples
In New Zealand, the government has launched a national AI strategy and a public service AI framework, but lacks environmental disclosure requirements or a regulator to track energy use and emissions. Similarly, Australia's national AI plan focuses on improving public services, such as the National Film and Sound Archive's Bowerbird transcription tool and the Department of Veteran's Affairs' claims processing proof-of-concept. Both countries adopt a light-touch, principles-based regulatory approach, which the report warns may overlook the growing environmental costs of AI.
The UN emphasizes that the natural environment is foundational to the economy, culture, and wellbeing, and must be central to AI development. It calls for a rethinking of the AI innovation playbook, shifting focus toward a sustainable tech future that balances capability with environmental stewardship.
