The United Kingdom is currently navigating two competing national ambitions: achieving a decarbonized economy through renewable energy and establishing itself as a global AI superpower. However, a significant discrepancy in data between the government departments tasked with these goals suggests that the UK may not have a unified plan for managing the massive energy requirements of the digital age.
The Data Discrepancy
A major rift has emerged between the Department for Science, Innovation and Technology (DSIT) and the Department for Energy Security and Net Zero (DESNZ) regarding how much electricity AI datacentres will consume by 2030.
- DSIT’s Vision: In its “compute roadmap,” DSIT forecasts that the UK will require at least 6GW of AI-capable datacentre capacity by 2030 to transform the national computing ecosystem.
- DESNZ’s Forecast: In contrast, DESNZ’s projections for the entire “commercial services” sector—which includes datacentres—suggest an energy increase of only 528MW between 2025 and 2030.
To put this in perspective, DSIT’s requirement for AI alone is more than ten times higher than the total growth DESNZ expects for the entire commercial services sector. Furthermore, a single “AI growth zone” (a hub designed to attract investment) is expected to require roughly 500MW—nearly the same amount of energy DESNZ allocates to the entire sector’s growth for the next five years.
Why This Matters: The Environmental Impact
This misalignment is not merely a bureaucratic error; it has profound implications for the UK’s ability to meet its international climate targets.
DESNZ is responsible for delivering the UK’s carbon budget. If the energy demands of AI are significantly underestimated, the government may find itself unable to meet its net-zero commitments. This raises critical questions about whether the rapid expansion of AI infrastructure is being planned with environmental sustainability in mind, or if “magical thinking” is being applied to the energy needs of Big Tech.
Shifting Numbers and Accountability
The controversy has already led to significant revisions in official government documentation. Following inquiries regarding the plausibility of their data, DSIT recently updated its projected carbon emissions for the AI sector:
- Initial Figures: DSIT originally projected emissions between 0.025m and 0.142m tonnes of CO2e—a negligible amount representing less than 0.05% of total UK emissions.
- Revised Figures: Following scrutiny, DSIT updated these figures to a range of 34 to 123 MtCO₂, which accounts for roughly 0.9% to 3.4% of the UK’s projected total emissions over the next decade.
While DSIT argues that grid decarbonization will help keep emissions toward the lower end of this range, the massive jump in projected impact highlights how much the initial assessments underestimated the carbon footprint of AI.
Expert Concerns
Critics argue that this lack of coordination reveals a vulnerability in government planning.
“The government’s cluelessness over the environmental impact of datacentres would be laughable, if it weren’t so alarming,” said Tim Squirrell, head of strategy for the NGO Foxglove.
Cecilia Rikap, a researcher at University College London, suggested that the discrepancy points to a deeper issue of influence, noting that the misalignment could suggest that large technology corporations exert undue influence over government policy and projections.
Conclusion
The massive gap between the UK’s AI ambitions and its energy modeling creates a strategic blind spot. Unless the government synchronizes its technological goals with its climate commitments, the push for AI leadership may directly undermine the transition to a green economy.
