Officials hugely underestimated impact of AI datacentres on UK carbon emissions

The UK government has significantly underestimated the carbon emissions associated with artificial intelligence (AI) datacentres, revising its figures upward by more than 100 times. New data indicates that AI-related energy consumption could produce up to 123 million tonnes of carbon dioxide (CO₂) over the next decade, equivalent to the emissions of approximately 2.7 million people. This stark increase replaces an earlier estimate that projected a maximum of just 0.142 million tonnes of CO₂ annually. The revision raises serious concerns about the environmental impact of AI infrastructure amid growing urgency to reduce global emissions. The updated figures were published as part of a revision to the UK’s “compute roadmap,” a government strategy aimed at establishing a world-class AI ecosystem to drive economic growth. However, AI datacentres require substantially more electricity than traditional data storage facilities, and much of this energy is still derived from fossil fuels. According to the Department for Science, Innovation and Technology (DSIT), the carbon footprint of the planned AI expansion could account for between 0.9% and 3.4% of the UK’s total projected emissions from 2025 to 2035. The lower end of this range assumes improvements in AI efficiency and faster decarbonisation of the national energy grid. The revision followed scrutiny from independent watchdog Foxglove and the Carbon Brief news site, which challenged the government’s initial estimates as being unrealistically low. Foxglove’s head of strategy, Tim Squirrell, highlighted the tension between the UK’s legally binding net zero target for 2050 and the rapid expansion of energy-intensive AI datacentres, which could potentially double the country’s electricity consumption if left unchecked. Climate campaigners have expressed alarm, warning that prioritising AI development without addressing its environmental costs risks exacerbating the climate crisis and squandering limited carbon budgets. This development underscores the complex challenge of balancing technological advancement with sustainability goals. As the UK government invests heavily in AI infrastructure, including a recent £500 million fund, the environmental implications of such growth are coming under increased scrutiny. The revised emissions estimates serve as a critical reminder of the need for integrating energy efficiency and clean energy solutions into the future of AI deployment to avoid undermining national and global climate commitments.