Energy Minister Michael Shanks has issued a clear warning that the burgeoning Artificial Intelligence (AI) sector must actively contribute to lowering household energy bills, rather than exacerbating the strain on Britain's already stretched electricity network. Speaking at the UKAI Infrastructure Summit on Wednesday, Mr Shanks underscored that public backing for the swift expansion of AI infrastructure is contingent on tangible economic benefits for ordinary citizens.
“We need to deliver energy security here at home, but we need to bring down bills,” Mr Shanks stated, highlighting that affordability remains a top government priority. His comments come as AI developers increasingly compete for robust and reliable power supplies essential for running large data centres, adding considerable pressure to a national grid that is already struggling to expand its capacity to meet demand.
The government's own projections for carbon emissions linked to its AI infrastructure plans have recently seen a substantial upward revision. Officials now anticipate that AI data centres could generate between 34 million and 123 million tonnes of carbon dioxide over the coming decade. The higher estimate alone could account for approximately 3.4 per cent of Britain's total projected emissions between 2025 and 2035, depending on efficiency improvements and the pace of power system decarbonisation.
Evidence is mounting that insufficient grid capacity could impede investment in British AI. For instance, Nscale, a British company backed by Nvidia, is reportedly exploring alternative power solutions for its planned £2 billion data centre in Essex. This follows delays to its National Grid connection, which is not expected to be ready by the site's anticipated opening. More than 100 UK development projects are reportedly considering temporary on-site gas generation as they await permanent grid access, with some facing connection delays stretching up to a decade, despite data centres being central to the government's strategy for building domestic AI capabilities.
Mr Shanks proposed that strategically locating data centres near renewable energy generation could alleviate some of the pressure. He explained that Scotland, for example, often produces more renewable power than the network can transmit south, leading to wind farms reducing output during periods of grid congestion. Consumers currently bear the cost of these curtailments through their bills. By placing data centres in such areas, they could absorb excess energy, thereby helping to reduce overall costs for consumers. He also suggested that increased industrial demand could help spread the cost of the electricity system beyond domestic customers, who currently shoulder a significant portion of the expense. Furthermore, the minister highlighted AI's potential to optimise the existing network, reducing the reliance on entirely new infrastructure like pylons and substations.