TLDR: A new study from Cornell University reveals that the rapid expansion of AI data centers could lead to a significant surge in carbon emissions and water consumption by 2030. However, the research provides a ‘roadmap’ detailing three key strategies—smarter facility siting, accelerated grid decarbonization, and increased operational efficiency—that could collectively cut carbon emissions by 73% and water usage by 86% compared to worst-case projections.
The burgeoning growth of artificial intelligence (AI) infrastructure, particularly data centers, poses a substantial environmental challenge, according to a recent study by Cornell University researchers. Published in Nature Sustainability on November 10, 2025, the study projects a dire outlook if current trends continue, but also offers actionable solutions.
Projected Environmental Toll by 2030 (Worst-Case Scenario):
If the current rate of AI growth persists, the study forecasts that by 2030, U.S. AI data centers could annually emit between 24 and 44 million metric tons of carbon dioxide. This is equivalent to adding 5 to 10 million cars to American roadways. Concurrently, water consumption is projected to reach 731 to 1,125 million cubic meters per year, matching the annual household water usage of 6 to 10 million Americans. These figures, the researchers warn, would render the AI industry’s net-zero emissions targets unattainable. Furthermore, AI electricity use is expected to surge 7 to 17 times, carbon emissions 2.5 to 7 times, and water use 6 to 13 times between 2024 and 2030.
A Three-Pronged ‘Roadmap’ for Sustainability:
Led by Fengqi You, the Roxanne E. and Michael J. Zak Professor in Energy Systems Engineering at Cornell, the research team has outlined a comprehensive strategy to mitigate these impacts, potentially reducing carbon emissions by 73% and water consumption by 86% against worst-case scenarios. “Artificial intelligence is changing every sector of society, but its rapid growth comes with a real footprint in energy, water and carbon,” stated Professor You. “Our study is built to answer a simple question: Given the magnitude of the AI computing boom, what environmental trajectory will it take? And more importantly, what choices steer it toward sustainability?”
1. Smarter Facility Siting:
The location of data centers is a decisive factor. Many existing and planned data clusters are situated in water-stressed regions like Nevada and Arizona, or in rapidly expanding hubs such as northern Virginia, which can strain local resources. The study identifies Midwest and ‘windbelt’ states, including Texas, Montana, Nebraska, and South Dakota, as optimal locations due to their balanced carbon and water profiles. New York state also presents a favorable option, owing to its clean electricity mix of nuclear, hydroelectric, and growing renewable sources, especially when coupled with water-efficient cooling technologies. Relocating facilities to areas with lower water risk and improving cooling efficiency could halve water demands alone.
2. Accelerated Grid Decarbonization:
While crucial, decarbonizing the electrical grid alone is insufficient. Even in an ambitious high-renewables scenario, carbon dioxide emissions would only drop by approximately 15% by 2030, leaving a residual 11 million tons of CO2. This would necessitate an additional 28 gigawatts of wind or 43 gigawatts of solar capacity to achieve net-zero. Professor You emphasized, “Even if each kilowatt-hour gets cleaner, total emissions can rise if AI demand grows faster than the grid decarbonizes. The solution is to accelerate the clean-energy transition in the same places where AI computing is expanding.”
3. Increased Operational Efficiency:
Implementing advanced energy- and water-efficient technologies within data centers is vital. This includes innovations such as advanced liquid cooling systems and improved server utilization. These measures could contribute an additional 7% reduction in carbon emissions and a 29% decrease in water use, leading to a total water reduction of 32% when combined with other strategies.
Professor You concluded, “There isn’t a silver bullet. Siting, grid decarbonization and efficient operations work together – that’s how you get reductions on the order of roughly 73% for carbon and 86% for water.” He added, “This is the build-out moment. The AI infrastructure choices we make this decade will decide whether AI accelerates climate progress or becomes a new environmental burden.”
Research Background:
Also Read:
- Charting a Sustainable Path for Artificial Intelligence: A Deep Dive into Green AI Practices
- Agentic AI Transforms Supply Chain Sustainability and Efficiency
The three-year study involved compiling extensive financial, marketing, and manufacturing data, alongside location-specific information on power systems and resource consumption. The team also utilized AI to bridge data gaps. The first author of the study is doctoral student Tianqi Xiao. Collaborators included researchers from the KTH Royal Institute of Technology (Sweden), Concordia University (Canada), and RFF-CMCC European Institute (Italy). The project received support from the National Science Foundation and the Eric and Wendy Schmidt AI in Science program.
