06 Jun 2026
Tired Earth
By The Editorial Board
Artificial intelligence is often portrayed as an invisible technology—software running quietly in the background, powering chatbots, search tools, and automated systems. Yet behind every AI-generated response lies a vast physical infrastructure that consumes electricity, water, land, and raw materials on an unprecedented scale.
A recent United Nations report highlights a growing challenge: as AI adoption accelerates worldwide, the environmental demands of the data centers that support it are increasing just as rapidly. The findings suggest that the future of AI will not be shaped solely by advances in computing power, but also by the availability of natural resources.
For many users, AI feels intangible. In reality, every query submitted to an AI system is processed in data centers filled with specialized hardware operating around the clock. These facilities require enormous amounts of electricity to run servers and significant quantities of water to prevent equipment from overheating.
As AI models become larger and more widely used, the pressure on this infrastructure continues to grow. According to the UN analysis, global demand from data centers is expected to rise sharply during the remainder of the decade, driven largely by AI-related workloads.
Water has emerged as one of the least understood aspects of the AI boom.
Data centers often rely on cooling systems that dissipate heat through evaporation. While this approach improves operational efficiency, it can require substantial volumes of freshwater. Water is also indirectly consumed through electricity generation, creating an environmental footprint that extends beyond the facilities themselves.
The UN report warns that water use associated with AI and data-center operations could reach levels that compete with the needs of large populations, particularly in regions already facing water stress. As demand for computing capacity expands, questions about where facilities are built and how they are cooled are likely to become increasingly important.
Energy consumption presents a similar challenge.
The expansion of AI services is contributing to a rapid increase in electricity demand from data centers worldwide. Training advanced AI models requires enormous computational resources, while serving billions of user requests each day adds further pressure to electrical grids.
The report projects that data-center electricity consumption could increase dramatically by 2030. Such growth would place AI infrastructure among the most energy-intensive sectors of the digital economy, forcing governments and utilities to consider how future demand will be met.
The issue is not only the quantity of energy consumed but also its source. Facilities powered primarily by fossil fuels generate significantly higher greenhouse-gas emissions than those supplied by renewable or low-carbon electricity.
Public discussions about technology often focus on carbon emissions, but the UN report argues that the environmental footprint of AI is broader.
Expanding data-center networks requires land for buildings, power infrastructure, and supporting facilities. Manufacturing the hardware that powers AI systems consumes minerals and industrial resources, while retired equipment contributes to growing volumes of electronic waste.
The report notes that environmental trade-offs are frequently overlooked. In some cases, solutions that reduce carbon emissions may increase pressure on water resources or land use. As a result, evaluating AI's sustainability requires a more comprehensive approach than measuring emissions alone.
One area attracting particular attention is AI-generated multimedia content.
Creating images, audio, and especially video typically requires far more computational power than simpler tasks. As these applications become mainstream, they may significantly increase the resource demands associated with AI services.
Researchers cited in the UN report describe this trend as an emerging environmental concern, emphasizing that not all AI activities have the same footprint. The type of content being generated can substantially influence energy consumption.
The report does not argue against AI development. Instead, it calls for stronger governance and greater transparency across the industry.
Among its recommendations are improved reporting of environmental impacts, greater efficiency in system design, responsible management throughout the technology lifecycle, and stronger international cooperation. The report also stresses the importance of environmental justice, ensuring that the costs of AI expansion do not fall disproportionately on communities with limited resources.
Responsibility, the authors argue, extends beyond technology companies. Policymakers, researchers, hardware manufacturers, investors, businesses, media organizations, and end users all have a role to play in shaping a more sustainable digital future.
The commercial potential of AI is enormous, with analysts forecasting dramatic market growth over the next decade. Yet the UN report suggests that the industry's long-term success may depend on how effectively it addresses its environmental footprint.
Artificial intelligence promises major advances in productivity, healthcare, science, and education. However, achieving those benefits sustainably will require acknowledging that AI is not merely software. It is a resource-intensive industrial ecosystem whose growth depends on energy networks, water supplies, physical infrastructure, and responsible governance.
The challenge for the coming decade will be ensuring that technological progress does not outpace the planet's capacity to support it.
The best of Tired Earth delivered to your inbox
Sign up for more inspiring photos, stories, and special offers from Tired Earth
By signing up for this email, you are agreeing to news, offers, and information from Tired Earth. Click here to visit our Privacy Policy.
Comment