UN asks AI companies to reveal full environmental impacts

The head of the United Nations has launched an initiative aimed at holding artificial intelligence companies accountable for their exploding environmental impacts, including their carbon emissions, the amount of water and land used for data centres, and the energy they consume.

During a speech at London Climate Action Week on Tuesday, António Guterres noted that AI can accelerate climate solutions, among other key challenges, and said its potential must be harnessed.

“But AI is also hungry for land, water and power,” he emphasised, adding that the data centres needed to run AI models already consume more electricity than most countries. 

The UN Secretary-General repeated a call he first made in July 2025 for all big AI companies to commit to power every data centre with renewable energy by 2030. 

The head of the United Nations has launched an initiative aimed at holding artificial intelligence companies accountable for their exploding environmental impacts, including their carbon emissions, the amount of water and land used for data centres, and the energy they consume.

During a speech at London Climate Action Week on Tuesday, António Guterres noted that AI can accelerate climate solutions, among other key challenges, and said its potential must be harnessed.

“But AI is also hungry for land, water and power,” he emphasised, adding that the data centres needed to run AI models already consume more electricity than most countries. 

The UN Secretary-General repeated a call he first made in July 2025 for all big AI companies to commit to power every data centre with renewable energy by 2030. 

Some tech firms have announced they are sourcing or building out clean energy to run their hubs, but growing power demand is also contributing to gas-fired generation in the US, according to data from Global Energy Monitor.

The International Energy Agency (IEA) estimates that data centres are set to more than double the emissions from the electricity they use between 2024 and 2030 in a high-growth scenario. But AI’s use could lead to far larger reductions in the energy sector through efficiency gains if adopted widely.

‘No more hidden costs’

Proposing the new “AI Environmental Transparency Initiative” on Tuesday, Guterres also urged big AI firms companies to measure and publicly disclose the full environmental impact of their systems, including their carbon, water, and land footprints.

“No more hidden costs. No more shifting the burden onto those least able to bear it. It is time to come clean,” he said in a major speech on responding to the world’s twin climate and energy crises. “If AI is to help build a better future, it must be honest about what it costs us now.”

A report issued earlier this month by the UN University Institute for Water, Environment and Health noted that most current assessments of AI’s environmental cost focus on carbon emissions from training models. But, it added, this misses a substantial part of the picture. 

Every kilowatt-hour of electricity for AI also carries a water footprint, from cooling and generation, and a land footprint, from infrastructure and supply chains, it said. 

Explainer: Will AI data centres make or break the energy transition?

The report estimated that AI data centres globally could consume 945 terawatt-hours of electricity annually by 2030 –  more power than all but five countries and roughly twice France’s 2025 consumption.

Offsetting this carbon footprint by 2030 would require growing some 6.7 billion trees over 10 years, it calculated. Producing power for the data centres would consume water equal to the basic needs of 1.3 billion people in sub-Saharan Africa for a year and take up land of more than 14,500 square kilometers, roughly twice the Jakarta metropolitan area.

The European Union said earlier this month it will develop minimum energy-efficiency standards for both new and existing data centres, with a “needs assessment” ​due by 2027, Reuters reported. It’s also planning ⁠a sustainability label for data centres, covering criteria including water use and clean energy supply – but that has been delayed.    

US community push-back 

Asked after his speech what the response had been, the UN chief said “we’ll see”, without giving more details.

But, he argued that, in his view, the push for transparency “is perfectly reasonable and even positive for the AI industry, because eventually some people will say that they consume much more than they really do”. “I think the truth is essential,” he added. 

Concerns about the environmental impacts of AI and the infrastructure needed to run the technology have led to growing opposition in some communities, especially in the US.

This month, Monterey Park in Los Angeles County was the first city in the United States to enact a citywide prohibition on data centres through a voter-approved ballot measure. The developers behind a proposed centre in the area had already pulled the project in April amid an increasingly hostile local environment and regulatory uncertainty.

The vote that stopped a data center: US communities query resource-hungry AI

According to nonprofit Data Center Watch, around $64 billion-worth of data centre projects nationwide were delayed or blocked between May 2024 and March 2025 as communities pushed back against them. 

Industry lobby groups argue that data centres can provide economic benefits in their host communities. According to the US-based Data Center Coalition, which represents big operators and developers, data centres generate tax revenue, support construction and technical jobs, and provide infrastructure needed for cloud computing, scientific research and AI development.

The industry has also challenged claims that data centers necessarily raise electricity costs for households.

Force for good?

The UN chief said benefits can be few in the places that are home to the data centre, while “communities are often left in the dark about the environmental impact of the infrastructure rising around them”.

Guterres said companies have an “obligation” to be clear and open about the services they are offering but also the level of resources they require. 

“Transparency is essential for the decisions that communities must make – and transparency is essential even for the future of artificial intelligence, and to make sure that artificial intelligence is essentially a force for good,” he told an audience of climate professionals in London

A senior UN official told journalists ahead of Tuesday’s announcement that the AI industry has started to talk about and disclose some of their impacts, but those efforts are not yet comprehensive enough. 

The hope is that the new initiative will “encourage the industry to come together and take further action on it”, the official said.