A new analysis estimates that artificial intelligence activity in 2025 generated up to 80 million tons of CO2—roughly matching New York City’s annual emissions—and consumed about 765 billion liters of water, exceeding global bottled-water demand. The research, published in Patterns by Digiconomist founder Alex de Vries-Gao, argues AI-specific use now accounts for more than 8% of worldwide aviation emissions and calls for tighter corporate disclosure on environmental impacts. The International Energy Agency says AI-driven data centers are set to more than double electricity consumption by 2030, with the largest facilities rivaling usage by millions of homes. Advocates say communities are bearing the costs as tech firms accelerate a “hyperscale” buildout, citing UK and India projects that could lock in higher emissions through backup diesel generation. Google says it cut 2024 data center energy emissions by 12% with cleaner power but acknowledges slower-than-needed deployment of carbon-free energy is a constraint. The study underscores growing pressure on tech companies to account for full lifecycle energy and water use tied to AI services.
Related articles:
Making AI Less “Thirsty”: Uncovering and Addressing the Secret Water Footprint of AI Models
Carbon Emissions and Large Neural Network Training
