The UK government faces a major challenge: promoting clean energy while also aiming to make the country a superpower in artificial intelligence (AI). However, there appears to be serious inconsistencies between the departments responsible for these goals. The Department of Science, Innovation and Technology (DSIT) and the Department for Energy Security and Net Zero (DESNZ) have vastly different estimates regarding the future electricity demand of AI data centers.

DSIT estimates that by 2030, the electricity demand of AI data centers will reach 6 gigawatts, while DESNZ believes this number will be less than one-tenth of that. This gap has drawn attention from the non-governmental organization Foxglove, whose strategy director, Tim Squirrell, said the government's ignorance of the environmental impact of data centers is shocking. Cecilia Ricap, a researcher at University College London, also pointed out that this discrepancy may stem from departmental capacity limitations or unrealistic expectations about large technology companies.

DESNZ is responsible for the UK's carbon budget plan, explaining how the government will meet its international climate goals. Foxglove had submitted an environmental impact assessment request to DESNZ, asking how the growth of AI data centers could be incorporated into carbon emission projections. DESNZ responded that researchers could refer to the overall energy consumption forecasts of the commercial services sector, without making specific predictions about the growth of data centers.

According to this forecast, the energy consumption of the entire industry will increase by 528 megawatts between 2025 and 2030, equivalent to the electricity needs of about 170,000 households. This projected value is far lower than ten times the power required by AI data centers mentioned in DSIT's "UK Computing Roadmap." DSIT's report emphasizes that by 2030, the UK will need at least 6 GW of AI data center capacity.

During this period, DSIT seems to have revised its earlier AI data center emissions projections, increasing the emissions figures by more than a hundredfold. Initially, DSIT predicted carbon emissions from additional AI computing capacity to be between 0.025 and 0.142 million tons, which was later updated to between 34 and 123 million tons, accounting for 0.9% to 3.4% of the UK's estimated total emissions.