The International Energy Agency estimates that if AI is used across transport, industry and buildings, it could save 13.5 exajoules of energy by 2035 – about the same amount Indonesia uses in a year. This shows how AI can directly lower overall electricity consumption.

Siddharth Singh, an AI expert at the International Energy Agency, says AI can give electricity operators very local forecasts of cloud cover, sunlight and wind. Those forecasts let grids predict renewable output more accurately and cut the amount of curtailment. WattTime, an environmental‑tech nonprofit, uses the same idea to schedule washing machines, electric vehicles and other devices when clean solar power is plentiful.

Florian Pappenberger, director‑general of the European Centre for Medium‑Range Weather Forecasts, explains that AI has made weather prediction much faster. The centre can now produce forecasts in two years that would have taken ten years before, and it does so at lower cost. Faster, cheaper forecasts also improve early‑warning systems for extreme weather.

Satellites combined with AI in the Climate TRACE initiative track greenhouse‑gas emissions from facilities worldwide, creating detailed inventories. The project, co‑founded by Gavin McCormick, provides emission data that cities can use for climate plans. At the same time, Maria Joao Sousa, executive director of Climate Change AI, warns that oil and gas companies also use AI to analyse geological data and increase fossil‑fuel production, which harms the climate.

AI therefore offers powerful tools to cut emissions and make renewable energy work better, but it can also accelerate fossil‑fuel extraction. Careful policies are needed to ensure the technology is used for climate benefits rather than harms.