Could AI End Up Drinking More Water Than 8 Billion People? Artificial intelligence is often promoted as a tool that will become more efficient over time, reducing its environmental footprint as technology advances. However, a new United Nations report argues that this assumption may be dangerously misleading.
According to the report, the rapid expansion of AI systems could dramatically increase global demand for electricity, water, and land over the next decade, potentially creating environmental pressures on a scale few policymakers currently appreciate.
The warning comes as governments and technology companies race to build increasingly powerful AI models, requiring vast networks of energy-intensive data centers around the world.
AI’s Environmental Footprint Is Growing Fast
The UN estimates that by 2030, AI-related infrastructure could consume nearly 3% of the world’s total electricity supply, roughly double current levels.
Such growth would place AI among the largest consumers of energy globally, raising concerns about its contribution to climate change and resource depletion.
Researchers behind the report argue that the environmental costs of AI extend far beyond electricity usage. Massive data centers require enormous amounts of water for cooling systems, as well as large areas of land and significant quantities of construction materials.
If current trends continue, AI-powered infrastructure could become one of the most resource-intensive sectors of the global economy.
The Water Challenge
Perhaps the report’s most striking finding concerns water consumption.
By the end of the decade, data centers supporting AI technologies could require approximately 9.3 trillion liters of water annually for cooling operations.
That figure exceeds the estimated amount of drinking water needed by the world’s population over the same period.
As climate change intensifies droughts and water shortages across many regions, experts warn that growing competition between technological infrastructure and basic human needs could become a major policy challenge.
Water stress is already affecting numerous countries, and the rapid expansion of AI facilities may place additional pressure on vulnerable ecosystems and communities.
The Jevons Paradox: Why Efficiency May Not Help
One of the report’s central arguments revolves around an economic concept known as the Jevons paradox.
The theory suggests that improvements in efficiency often lead to higher overall consumption rather than lower consumption.
The concept dates back to the 19th century when economist William Stanley Jevons observed that improvements in coal-use efficiency in England did not reduce coal demand. Instead, lower costs encouraged broader adoption and increased overall consumption.
The UN believes a similar pattern could emerge with AI.
As AI models become faster, cheaper, and more efficient, organizations may deploy them across a wider range of industries and applications. While individual systems may require fewer resources, the overall scale of usage could rise dramatically.
In other words, efficiency gains alone may not reduce AI’s environmental impact if demand continues growing at an even faster pace.
Electricity Demand Approaching National Levels
The report highlights how quickly data center electricity consumption is already increasing.
In 2025, global data centers consumed approximately as much electricity as Saudi Arabia, one of the world’s largest energy consumers.
If current projections prove accurate, energy demand from AI infrastructure could double by 2030.
Such growth would require major investments in power generation, grid expansion, and energy storage systems.
Without a substantial increase in renewable energy deployment, much of this demand could be met through fossil fuels, increasing greenhouse gas emissions.
A Significant Carbon Footprint
The climate implications are equally concerning.
According to the report, the carbon emissions associated with projected AI-related electricity demand would be so large that offsetting them would require roughly 6.7 billion trees grown over a ten-year period.
Environmental researchers stress that while AI can support climate solutions in areas such as energy optimization and scientific research, those benefits must be weighed against the emissions generated by the infrastructure powering these systems.
Land Use Pressures Also Rising
Beyond energy and water, AI expansion is expected to require significant land resources.
The UN estimates that future data center growth could occupy an area nearly ten times larger than Mexico City, creating additional environmental and urban planning challenges.
Large-scale facilities often require extensive transportation links, power infrastructure, and cooling systems, further increasing their ecological footprint.
A Call for Responsible AI Development
Rather than arguing against AI innovation, the report advocates a more sustainable approach to development.
The UN proposes a framework built around several key principles:
- Transparency in environmental reporting
- Efficiency-focused system design
- Fair and equitable resource allocation
- Lifecycle accountability for hardware and infrastructure
- International cooperation
- Sustainable long-term deployment strategies
Researchers argue that environmental considerations should become a core part of AI policy and investment decisions rather than an afterthought.
The Road Ahead
Artificial intelligence is expected to transform industries ranging from healthcare and education to finance and manufacturing. Yet the UN warns that the technology’s environmental consequences cannot be ignored.
The coming decade will likely determine whether AI becomes a model for sustainable technological progress or a major new source of pressure on the planet’s energy, water, and land resources.
As demand for AI continues to surge, policymakers, technology companies, and consumers may face a difficult question: Can the benefits of artificial intelligence be achieved without overwhelming the natural resources that support modern society? Can SpaceX Justify a $1.77 Trillion Price Tag? | Maya
