A new United Nations University (UNU) report warns that the rapid expansion of artificial intelligence could place significant pressure on global water, land, and energy systems.

While AI is often assessed through its carbon emissions, the study highlights broader environmental impacts, including large-scale electricity consumption, water use for cooling, land requirements for infrastructure, and rising electronic waste.

Rising Electricity Demand from AI Infrastructure

Data centers, which power artificial intelligence systems worldwide, are projected to consume up to 945 terawatt-hours of electricity annually by 2030. This is nearly three times the combined annual electricity use of Pakistan, Bangladesh, and Nigeria, countries with a total population of more than 650 million people.

The report notes that this figure represents only part of AI’s environmental footprint. Beyond carbon emissions, every unit of electricity used in data centers also carries a water footprint for cooling and energy production, as well as a land footprint linked to power generation infrastructure and supply chains.

According to the UNU study, AI-related electricity demand alone could account for nearly 3% of global electricity use by 2030, with emissions comparable to those of the United Kingdom in 2025.

Water and Land Pressures Often Overlooked

The report highlights that environmental assessments often focus narrowly on greenhouse gas emissions, particularly those associated with training large AI models, while overlooking other key impacts.

By the end of the decade, AI-related water consumption could equal the annual domestic needs of 1.3 billion people, while its land footprint may exceed 14,500 square kilometers, roughly twice the size of the Jakarta metropolitan area.

The study also notes that so-called “green” solutions in one area may increase pressure elsewhere. For example, shifting to certain renewable energy sources can reduce carbon emissions but significantly raise water and land use requirements.

Daily AI Usage Drives Most Energy Consumption

While public debate has largely focused on the energy required to train AI models, the report finds that daily usage accounts for around 80-90% of total energy demand.

One widely used AI service is estimated to process around 2.5 billion prompts per day, consuming hundreds of gigawatt-hours of electricity annually.

Energy use also varies significantly depending on the task. Generating a single AI image may require more than 1,000 times the energy needed for simple text classification, while video generation demands even more resources.

The report warns that efficiency gains alone are unlikely to reduce overall consumption due to a rebound effect, where lower costs and improved performance lead to increased usage.

Uneven Global and Environmental Impacts

The environmental burden of AI infrastructure is not evenly distributed. While AI benefits are global, its costs are often concentrated in specific regions.

In some countries, data centers already account for a significant share of national electricity consumption. In others, expanding facilities place additional pressure on water resources, sometimes in areas already affected by drought.

The report also raises concerns about electronic waste, estimating that AI infrastructure could generate up to 2.5 million tonnes of e-waste annually by 2030, much of which may end up in lower-income countries with limited disposal capacity.

In addition, demand for critical minerals used in AI hardware raises concerns about environmental degradation and social inequality in extraction regions.

Digital Inequality and Infrastructure Concentration

The UNU study highlights a growing imbalance in global AI infrastructure. More than 90% of AI-specialized computing capacity is concentrated in just two countries, the United States and China, while over 150 countries lack significant domestic AI infrastructure.

This imbalance limits access to technological benefits and raises concerns about environmental justice, as some countries bear resource and environmental costs without receiving proportional economic gains.

Towards a Responsible AI Framework

Despite its findings, the report does not call for limiting AI development but instead urges more responsible and sustainable governance of the sector.

It proposes a “responsible AI ecosystem” based on transparency, efficiency by design, lifecycle responsibility, equity, global cooperation, and sustainable use.

Governments are encouraged to integrate AI infrastructure into long-term planning for energy, water, and land use. Companies are urged to design systems that minimize resource consumption, while users are encouraged to consider lower-impact applications where possible.

The report concludes that the future of AI will depend not only on technological progress but also on policy and governance decisions made today.