Artificial intelligence is driving a hidden environmental crisis that affects 340 million people worldwide, according to modern analysis by Les Numériques. The report highlights how the rapid expansion of AI infrastructure, particularly data centers powering large language models and machine learning systems, is consuming vast amounts of electricity and water, often in regions already facing resource scarcity.
The surge in demand for AI computing has led to a significant increase in energy use, with global data center electricity consumption projected to double by 2026, reaching levels comparable to the total power usage of countries like Japan or Germany. Much of this growth is concentrated in areas where cooling systems rely heavily on freshwater, intensifying pressure on local supplies.
In regions such as the American Southwest, parts of India, and northern China, AI-driven data centers are competing with agriculture and residential needs for limited water resources. Some facilities use millions of liters of water per day for cooling, a figure that scales rapidly as more companies deploy AI at scale.
Beyond water, the carbon footprint of AI training runs is substantial. A single large model training session can emit as much carbon dioxide as five cars over their lifetimes, and with thousands of models being developed and retrained annually, the cumulative emissions are becoming a material factor in global tech-sector pollution.
Whereas tech firms have made public commitments to sustainability, including pledges to use renewable energy and improve efficiency, the report notes that actual progress varies widely. Many AI operations remain dependent on fossil-fuel-powered grids, especially in regions where clean energy infrastructure lags behind demand.
The analysis warns that without coordinated action — including stricter efficiency standards, greater transparency in resource use, and investment in sustainable cooling technologies — the environmental toll of AI could undermine broader climate goals and exacerbate inequalities in access to water and power.
For businesses and investors, the findings signal growing regulatory and reputational risks tied to AI expansion. As governments begin to scrutinize the ecological impact of digital infrastructure, companies may face pressure to disclose AI-related resource consumption alongside financial metrics, similar to emerging ESG reporting requirements.
Les Numériques concludes that the AI boom’s environmental side effects are no longer negligible externalities but systemic challenges requiring immediate attention from policymakers, industry leaders, and the tech community at large.
