Artificial intelligence is rapidly becoming one of the most resource-intensive technologies on the planet, according to new research highlighting the sector’s growing environmental footprint. In 2025 alone, AI systems are estimated to have consumed electricity and water on a scale comparable to a major global metropolis, intensifying concerns about emissions, transparency, and long-term sustainability.
Researchers estimate that AI-related computing generated up to 80 million metric tons of carbon dioxide emissions this year – roughly equivalent to the annual emissions of New York City. At the same time, global data centers supporting AI workloads are believed to have used as much as 760 billion liters of water, primarily for cooling high-performance servers.
AI’s Resource Appetite Is Accelerating
The surge in energy and water use is being driven by explosive growth in large language models, image generators, and AI-powered enterprise tools. Training and running these systems requires dense clusters of GPUs operating around the clock, pushing data centers to consume electricity at unprecedented rates.
Cooling has emerged as a critical bottleneck. As servers grow more powerful, operators increasingly rely on water-intensive cooling systems to prevent overheating. This has led to rising strain on local water supplies, particularly in regions already facing scarcity.
A major challenge for researchers is the lack of transparency. Large technology companies do not disclose detailed breakdowns of how much electricity or water their AI operations consume. As a result, scientists are forced to estimate usage based on partial data, infrastructure capacity, and industry benchmarks.
Experts warn that this opacity makes it difficult to assess the true environmental cost of AI or to design effective policy responses. As previously covered, calls are growing for standardized reporting requirements tied to energy use and emissions from data centers.
Means for Markets and Policy
The environmental impact of AI is increasingly intersecting with regulation, investment decisions, and public policy. Governments are under pressure to balance AI-driven economic growth with climate commitments, while utilities face rising demand from hyperscale data centers competing with households and industry.
For investors, sustainability concerns could become a material risk factor for major technology firms. Higher energy costs, water restrictions, or carbon pricing could raise operating expenses and complicate expansion plans for AI infrastructure.
Looking ahead, analysts say mitigation will require a combination of more efficient chips, alternative cooling technologies, renewable energy integration, and clearer disclosure standards. Without such measures, experts warn that unchecked AI expansion could trigger environmental stress severe enough to reshape how and where future data centers are built.
