Staff at leading AI companies and institutions are showing worsening signs of mental strain over the dangers of latest models, while climate implications of the sector's rapid growth mount. UN climate change head Simon Stiell told a New York audience that "energy-guzzling artificial intelligence is driving up planet-heating pollution from coal, oil and gas," urging AI companies to power data centres with renewables instead of fossil fuels. Former US vice-president Al Gore noted AI emissions "should be put in perspective," pointing out their greenhouse gas impact remains well below that of the world's uncovered landfills.

According to the International Energy Agency, data centre power usage currently accounts for well below 1 per cent of global emissions. However, aggregate power consumption of data centres has kept rising fast — by 17 per cent in 2025 alone, with the IEA projecting a doubling from that level by 2030. The rise of agentic AI — through which autonomous bots perform highly complex tasks without supervision — threatens to steepen that curve.

In a July paper, researchers at the Korea Advanced Institute of Technology found that an agentic AI system comparable to current commercial services used up to 136.5 times more energy per query than a basic chatbot session. Last month Zeke Hausfather, climate research lead at payments company Stripe, calculated his use of Anthropic's agentic AI system consumed about a tenth of the power demand of an average US household. Tech companies are now driving for mass consumer adoption, exemplified by Meta's launch of its "Muse" agent through WhatsApp, Instagram and Facebook.

Really large-scale adoption of agentic AI would have significant implications for the global energy system. The KAIST team modelled a scenario where daily AI agent requests rise to 13.7bn, roughly the current daily number of global searches.