AI’s Growing Footprint Raises Climate Concerns

A data centre in Ireland.

Artificial intelligence is not only contributing to growing greenhouse gas emissions but is also expanding its environmental footprint at a pace that could place increasing pressure on the planet’s natural resources.

Data centres — the global infrastructure powering AI systems — could consume up to 945 terawatt-hours of electricity annually by 2030, nearly three times the combined yearly electricity use of Pakistan, Bangladesh, and Nigeria, countries home to more than 650 million people.

However, electricity use is only part of the challenge. Every unit of energy consumed by data centres also carries a significant “water footprint” for cooling and power generation, alongside a “land footprint” linked to electricity production and supply chains.

Rethinking AI Sustainability

According to a new study by the UN University (UNU), AI-related water consumption could equal the basic annual domestic needs of 1.3 billion people by the end of the decade. Meanwhile, its land footprint could exceed 14,500 square kilometres — roughly twice the size of the Jakarta metropolitan area.

The report highlights a major gap in how AI’s environmental impact is assessed. While greenhouse gas emissions, particularly those associated with training large AI models, often receive the most attention, other environmental costs are frequently overlooked.

Researchers warned that solutions considered “green” in one area may create pressure elsewhere. For instance, some renewable energy sources may reduce carbon emissions but significantly increase water use and land demand.

Everyday AI Use Drives Demand

Public discussions have largely focused on the energy required to train advanced AI systems, but the study found that daily use accounts for roughly 80 to 90 per cent of total energy demand.

The scale is considerable. 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 consumption also varies sharply by task. Generating a single AI image can require more than a thousand times the energy needed for basic text classification, while video generation consumes even greater resources.

The report noted that efficiency gains alone are unlikely to solve the problem due to the “rebound effect,” where lower costs and improved performance encourage greater use, ultimately increasing overall resource consumption.

Uneven Environmental Burden

The environmental costs of AI infrastructure are not shared equally. While the benefits of AI are global, many of the environmental burdens are concentrated in specific regions.

In some countries, data centres already consume a large share of national electricity supplies, straining energy systems. Elsewhere, facilities are placing growing pressure on water resources, including in drought-prone areas.

The report also warned of a rising electronic waste challenge, with AI infrastructure expected to generate up to 2.5 million tonnes of e-waste annually by 2030. Much of this burden could fall on lower-income countries with limited disposal capacity.

At the same time, the extraction of critical minerals needed for AI hardware raises concerns over environmental degradation and social inequality in mining regions.

Call for Responsible AI

Despite the findings, researchers stressed that the study is not an argument against AI. Instead, it calls for urgent measures to ensure the technology develops within environmental limits.

The report proposes a framework for a “responsible AI ecosystem,” based on transparency, efficiency, equity, lifecycle responsibility, international cooperation, and sustainable use.

Governments are encouraged to incorporate AI infrastructure into energy, water, and land-use planning, while companies are urged to design systems that minimise resource consumption.

Ultimately, researchers said the future environmental impact of AI will depend not only on technological progress but also on the policy and governance decisions made today.