Gap between government estimates and reality

An analysis has shown that carbon emissions from data centers, the core infrastructure of the artificial intelligence (AI) 'revolution' promoted by the British government, may be hundreds of times higher than government estimates. There are growing concerns that dozens of data centers being built across the country could cause a surge in electricity demand, posing a serious threat to the UK's ability to meet its climate goals, according to analysis by climate media.

According to an analysis by the UK Department of Science, Innovation and Technology (DSIT), emissions from 11.2 gigawatts (GW) of AI-related computing power by 2035 are expected to amount to just 142,000 tonnes at most. However, this figure is only possible if the electricity supplied to data centers is almost completely decarbonized. One environmental activist dismissed this analysis as "nonsense."

Relying on gas power generation could be equivalent to Denmark's annual emissions

The core of the problem is the source of power supplied to data centers. Data centers can minimize emissions if they run solely on low-carbon electricity, but some industry officials argue that the use of gas-fired power must increase to realize the government's AI ambitions.

The analysis found that if the new data center were to receive a significant amount of power from gas-fired power, its carbon dioxide (CO2) emissions could be at least equivalent to Denmark's total annual emissions. This is in stark contrast to the figures presented by the government.

Soaring electricity demand and uncertain future

The UK currently has approximately 1.8GW of data center facilities, consuming more than 2% of the country's electricity. According to energy regulator Ofgem, companies have already committed to investing in 71 new data centers, which will require around 20GW of power. For reference, the UK's average electricity demand in 2025 is approximately 37GW.

Concerns about this increase in electricity demand are growing among lawmakers and environmental groups. Last year, the government's plan to meet its 2035 climate target stated that AI growth was "not factored into emissions projections", but Energy Secretary Ed Miliband said new data centers were included in modeling "increasing overall electricity demand".

When did this trend start?

The rapid expansion of data centers is in line with the AI boom that accelerated after the emergence of ChatGPT in 2022. As the introduction of AI by companies began in earnest from 2023, demand for large-scale computing power increased explosively, and from 2024 to 2025, as governments around the world entered the race for AI hegemony, expansion of data center infrastructure emerged as a core national strategy.

The British government is actively supporting data center expansion with the goal of turning the country into an ‘AI superpower.’ However, it is pointed out that in this process, the gap between the speed of clean energy transition and the speed of AI infrastructure expansion is gradually widening.

Future outlook [AI analysis]

Future emissions from UK data centers are likely to vary greatly depending on a number of variables, including the number of centers being built, the cleanliness of the power, and the timing of operation. The government is targeting a 'clean power system' by 2030, but additional demand from data centers may require a faster pace than the clean energy expansion currently underway.

If clean power growth fails to keep pace with increased data center demand, two scenarios are expected: First, it may be necessary to keep more gas generation on the grid. Second, some facilities have the potential to build their own on-site gas power generation facilities. In this case, it is expected that achieving the UK's 2035 climate target will become significantly more difficult.

Ultimately, how to balance the AI ​​revolution and climate goals is likely to emerge as a key challenge for UK energy policy. It is time to establish more realistic policies to reduce the gap between the government's optimistic outlook and reality.