AI Carbon Footprint: The Technology Race That Is Quietly Heating the Planet
A new UN report has revealed a troubling paradox: the AI revolution that many see as a climate solution is itself a fast-growing driver of energy consumption and emissions. Global data centres consumed approximately 448 terawatt-hours of electricity in 2025 — more than many large countries consume in total.
The Scale of AI Energy Appetite
Training a single large language model can emit as much CO2 as five cars driven for their entire lifetimes. Google, Microsoft, Amazon, and Meta have collectively committed to procuring over 50 gigawatts of new clean energy by 2030 specifically for their AI operations — but critics note this demand is growing faster than renewable capacity can be added.
AI Water Footprint: The Hidden Cost
Data centres consume an estimated 700 million litres of water per day globally in 2025 for cooling. Communities in water-scarce regions where AI infrastructure is being built have raised concerns about data centre water use competing with agricultural and municipal needs.
Can AI Be Part of the Solution
AI is being applied to climate solutions with real impact: optimising energy grids, accelerating materials discovery, improving climate models, and enabling precision agriculture. The UN report calls for mandatory energy disclosure from AI companies and a preference for renewable-powered data centres.
Tags: AI Carbon Footprint | Data Centres | Energy Consumption | Tech Emissions | Climate Solutions | Sustainable AI
Comments