UN Pushes for Greater Transparency on AI’s Environmental Footprint, Launches New Reporting Initiative

Speaking during London Climate Action Week, Guterres announced the U.N.’s AI Environmental Transparency Initiative and called on major AI developers to measure and report the environmental impacts associated with their data centers. He also urged companies to commit to powering those facilities entirely with renewable energy by 2030.

According to Reuters, which first reported the announcement, Guterres warned that the accelerating buildout of data centers supporting AI systems could have significant consequences for global resource consumption. Reuters reported that the U.N. chief said AI-related infrastructure could, by the end of the decade, consume more electricity than all but five countries and require enough water annually to meet the basic needs of approximately 1.3 billion people in sub-Saharan Africa.

“If AI is to help build a better future, it must be honest about what it costs us now,” Guterres said, according to Reuters.

The remarks come as governments, environmental organizations, and local communities increasingly examine the resource demands of AI development. Data centers, which power advanced AI models and cloud computing services, require large amounts of electricity for computing and substantial water supplies for cooling systems.

Recent research has reinforced those concerns. Earlier this month, researchers from the United Nations University Institute for Water, Environment and Health projected that data centers could double their power and water consumption by 2030 as AI adoption expands. The report warned that unchecked growth could strain electricity systems, increase pressure on water resources, and contribute to rising electronic waste.

Read more: Colorado Refines, Rather Than Retreats From, AI Regulation

The issue is also drawing greater attention from policymakers. While many major technology companies have announced voluntary climate commitments, environmental advocates have argued that disclosures regarding AI-specific energy use and resource consumption remain inconsistent. Analysts and academic researchers have noted that regulatory frameworks governing AI’s environmental effects remain fragmented, with relatively few jurisdictions requiring detailed reporting on model-level energy or water consumption.

The debate intersects with broader questions about competition and market concentration in the AI sector. A relatively small number of technology companies control much of the global infrastructure used to train and deploy advanced AI systems, including hyperscale cloud platforms and large-scale data center networks. As governments examine AI competition issues, regulators in several jurisdictions have also begun evaluating whether dominant firms should face greater transparency obligations regarding the environmental impacts of the infrastructure that underpins their market positions.

Guterres’ announcement did not include new regulatory requirements, and the U.N. initiative relies on voluntary participation. However, the proposal reflects growing international pressure on AI companies to provide more detailed information about how their technologies affect energy systems, water resources, and local communities.

Alongside the AI transparency initiative, Guterres launched a separate call to reduce methane emissions. Reuters reported that he urged fossil fuel companies to address methane leaks, end routine flaring practices, and adopt science-based standards for emissions management. Methane is estimated to account for roughly one-third of current global warming.

The U.N. secretary-general also said he plans to convene world leaders ahead of the COP31 climate conference in Turkey later this year to advance discussions on reducing dependence on fossil fuels and supporting what he described as a “just transition” toward cleaner energy sources. Reuters reported that the meeting is intended to help accelerate international climate commitments as governments remain off track to meet existing global emissions targets.

Source: Reuters