AI data centers could double global electricity use by 2030, UN report warns | carboncredit.io
esgnews.earthA new report from the United Nations University Institute for Water, Environment and Health (UNU-INWEH) warns that AI-driven data centers are on track to consume 945 terawatt-hours of electricity annually by 2030, roughly double the 448 TWh used in 2025. This would account for nearly 3 percent of all projected global electricity use. The report also highlights that the water footprint tied to this electricity consumption could reach 9.3 trillion liters per year, equivalent to the basic domestic water needs of all 1.3 billion people in Sub-Saharan Africa. The study notes that the inference phase of AI, where models respond to user prompts, accounts for 80 to 90 percent of total energy demand. Generating a single AI image requires 1,450 times more energy than basic text classification. The authors emphasize that efficiency gains alone will not solve the problem, as lower costs drive higher usage, a pattern known as the Jevons Paradox. They call for mandatory environmental transparency from tech companies and warn that renewable energy procurement for data centers can divert clean electricity from the public grid.