- Data centres powering AI could consume 945 terawatt-hours of electricity annually by 2030, nearly triple the combined electricity use of Pakistan, Bangladesh, and Nigeria.
- AI-related water consumption could equal the basic annual domestic needs of 1.3 billion people by the end of the decade.
- More than 90% of AI-specialised computing capacity is concentrated in the United States and China, while over 150 countries lack significant domestic AI infrastructure.
AI’s Resource Footprint Is Expanding Beyond Carbon
United Nations researchers are warning that the global race to scale artificial intelligence is creating a wider environmental burden than carbon reporting alone captures.
A new study from UN University says AI’s environmental costs are spreading across energy, water, land, minerals and waste systems. The report argues that current measurement practices remain too narrow. Most scrutiny focuses on greenhouse gas emissions, especially those linked to training large models.
That approach misses other pressures that can be just as material for governments, companies and investors. A solution that cuts carbon in one area may add stress elsewhere. For example, switching to some renewable energy sources can reduce emissions but increase water use or land demand.
For executives, the finding is clear. AI cannot be treated only as a digital productivity tool. It is also becoming a physical infrastructure sector with rising exposure to energy planning, water rights, land use and waste governance.
Daily AI Use Drives Most Energy Demand
Public debate has focused heavily on the electricity needed to train advanced AI models. The UNU study points to a different pressure point. It finds that day-to-day AI usage accounts for roughly 80% to 90% of total energy demand.
The scale is already large. One widely used AI service is estimated to process around 2.5 billion prompts a day. That activity consumes hundreds of gigawatt-hours of electricity each year.
Demand also varies by task. Generating a single AI image can require more than a thousand times the energy of simple text classification. Video generation consumes even more.
This matters for corporate AI strategies. Efficiency gains may lower the cost of each query, but they do not guarantee lower total resource use. The report points to the rebound effect, where cheaper and faster systems drive higher usage. Total demand can then rise, even when each individual task becomes more efficient.
Data Centres Add Water And Land Risk
Data centres are the backbone of AI, but their sustainability profile is not limited to electricity. Every unit of power used carries a water footprint linked to cooling and energy production. It also has a land footprint tied to power generation and supply chains.
The report estimates that data centres could consume 945 terawatt-hours of electricity each year by 2030. That is nearly triple the combined annual electricity use of Pakistan, Bangladesh, and Nigeria, countries with more than 650 million people in total.
UNU researchers also estimate that AI-related water consumption could equal the basic annual domestic needs of 1.3 billion people by the end of the decade. Its land footprint may exceed 14,500 square kilometres. That is roughly twice the size of the Jakarta metropolitan area.
These figures sharpen the governance challenge. In water-stressed regions, AI infrastructure can compete with households, agriculture and industry. In power-constrained markets, data centres can add pressure to grids already managing electrification and climate resilience.
RELATED ARTICLE: UN Report Unveils Dire Situation for Migratory Species: Urgent Call for Conservation
Environmental Costs Are Unevenly Distributed
The report warns that the benefits and burdens of AI are not shared equally. AI tools are global, but their infrastructure impacts are local.
In some countries, data centres already represent a significant share of national electricity demand. In others, new facilities are drawing on water supplies during drought conditions.
The e-waste challenge is also growing. AI infrastructure could generate up to 2.5 million tonnes of electronic waste each year by 2030. Much of that burden may fall on lower-income countries with limited capacity for safe disposal.
Critical minerals add another risk layer. AI hardware depends on materials extracted through global supply chains. Those supply chains can carry environmental damage and social inequities in mining regions.
The Digital Divide Is Also An Environmental Divide
The report links AI infrastructure growth to widening global inequality. More than 90% of AI-specialised computing capacity is concentrated in just two countries: the United States and China. More than 150 nations lack significant domestic AI infrastructure.
That imbalance limits economic participation. It also raises environmental justice concerns. Some regions may absorb the extraction, energy, water or waste impacts of AI without gaining equal access to its economic upside.
For investors, this creates a due diligence issue. AI exposure now sits across climate risk, supply chain governance, human rights, resource scarcity and digital inclusion.
Responsible AI Needs Resource Governance
UNU researchers do not argue against AI. The report calls for stronger governance so AI develops within planetary limits.
It sets out a responsible AI ecosystem based on transparency, efficiency by design, equity, lifecycle responsibility, global cooperation and sustainable use.
Governments are urged to integrate AI infrastructure into energy, water and land-use planning. Companies are encouraged to design systems that reduce resource consumption from the start. Users also have a role by choosing lower-impact applications where possible.
The core message for boards is practical. AI governance can no longer sit only with technology teams. It now belongs in sustainability, risk, procurement and capital allocation discussions.
As AI demand accelerates, the countries and companies that build responsibly will shape more than digital markets. They will influence how the next wave of infrastructure competes for water, power, land and minerals in a warming world.
The post UN Report Warns AI’s Water, Land and Climate Costs Are Rising Faster Than Governance appeared first on ESG News.