UK Government Revises AI Data Center Emissions Forecast Up 100-Fold to 123 Million Tonnes CO2

The UK government announced a substantial revision to its projections for greenhouse gas emissions stemming from artificial intelligence (AI) infrastructure. The new analysis indicates that carbon emissions from data centers supporting AI could be 100 times higher than previously estimated.

This revised assessment, detailed in an analysis accompanying the government’s “Compute Roadmap,” projects that UK AI-related greenhouse gas emissions will reach 123 million tonnes of CO2 between 2025 and 2035. This cumulative figure represents 3.4% of the UK's total projected emissions over that same decade. For context, an earlier forecast published in July 2025 estimated up to 0.142 million tonnes of CO2 from AI for the year 2035 alone.

Data centers, critical for housing the powerful computing systems that process AI algorithms, consume vast amounts of electricity. This energy demand directly contributes to carbon emissions if derived from fossil fuel sources, posing a growing challenge to climate targets as AI adoption expands. The previous modeling was withdrawn in March to incorporate new data and analysis, leading to the dramatic increase in projections.

This significant upward revision highlights the escalating energy demands of AI and its potential impact on national carbon reduction efforts. Governments and industry face the challenge of reconciling technological advancement with environmental sustainability. Attention will now turn to strategies for powering AI infrastructure with renewable energy sources and improving data center efficiency.

Watch for further policy responses from the UK government regarding energy supply for the burgeoning AI sector and how these updated figures will integrate into broader climate change mitigation plans.