Will Britain's AI revolution affect its net zero aims?

• Britain's AI growth intensifies energy demands
• Infrastructure struggles with renewable integration
• Government to explore nuclear option for AI power

Prime Minister Keir Starmer aims to turn Britain into an artificial intelligence (AI) superpower, but experts say that ambition risks undermining the country's green goals.

Starmer's plan, announced in January, includes the creation of zones for data centres and a "National Data Library" to train AI. It also aims to increase the use of AI in the public sector, including the National Health Service (NHS).

But AI uses much more energy and water than traditional data centres, so the government's push could exacerbate challenges in energy distribution and strain the national grid, endangering the government's goal of reaching zero carbon power by 2030, experts caution.

If Britain uses data centres at home as the back-end of its AI services, "we will be draining cheap and green electricity away from energy-poor communities," said Gopal Ramchurn, professor of AI at the University of Southampton.

He said AI presents "a new trade-off for society" between water and electricity for taxable businesses and power for the most vulnerable populations.

Here is what you need to know about Starmer's plans:

What is the new AI plan?

Starmer's "AI Opportunities Action Plan" comprises 50 recommendations, including unlocking data from the BBC, the British Library, the Natural History Museum and the National Archives, among others, to create a National Data Library to train AI. Anonymised data from the NHS will also be used.

The government wants a 20-fold increase in AI computing power under public control by 2030, which would include building a new supercomputer, and the creation of 'AI growth zones’ to speed up planning proposals and build more AI infrastructure.

Britain would also create its own AI models, piloting new AI schemes to be rolled out nationwide, and change copyright law to allow easier training.

The government intends to reform text and data mining regulations as well so they are comparable to the European Union. 

"The current uncertainty around intellectual property is hindering innovation and undermining our broader ambitions for AI, as well as the growth of our creative industries. This has gone on too long and needs to be urgently resolved," the plan said.

However, the Creative Rights In AI Coalition, a group that represents creative industries, said the "only uncertainty is around who has been using the UK's creative crown jewels as training material without permission and how they got hold of it, making transparency provisions vital".

What could this mean for the electricity grid?

AI data centres require huge amounts of power to run. A ChatGPT query, for example, uses 10 times more energy than does a Google search. 

The head of Brtain's National Grid, John Pettigrew, has predicted demand by commercial data centres will increase six-fold between now and 2035 and has acknowledged that the grid will need to be revamped to meet that challenge.

Ramchurn told Context that Britain does not have the transmission or distribution network, which would include substations and reinforced cables, to ferry power to the right points or enough batteries or smart homes to store that energy.

Even without new power demands, Britain will need to improve its national infrastructure to meet its net zero goals, experts said.

Sugandha Srivastav, a lecturer at the Smith School of Enterprise and the Environment at the University of Oxford, said, however, there was a lack of will among citizens to build more infrastructure.

"So much new infrastructure will have to be built that this is nothing short of the next industrial revolution, the paradigm shift that will move the entire system to clean production," she told Context.

"Instead, we have a narrative centered on NIMBYism (those arguing 'Not In My Back Yard') which is not only economically damaging but also environmentally disastrous because it means a continued reliance on imported fossil fuels."

How could Starmer's AI plan affect net zero ambitions?

Construction and energy demand from data centres could exacerbate existing problems caused by Britain's lack of energy storage and could make the country more reliant on 'dirty' power like gas and oil.

Wind power was Britain's largest source of electricity in 2024 at around 40%, topping gas-fired power plants for the first time. But because the grid is unable to store power effectively, the government still has to pay wind companies to curtail their output at times and rely instead on oil and natural gas.

For example, recent cold snaps ramped up demand for gas, which heats about 75% of the country's homes.

Projections by Loughborough University in December suggested energy demand from the digital data sector could outpace the growth of renewable energy capacity.

Decarbonisation also relies upon replacing gas boilers with electric heat pumps and the greater use of electric vehicles - all of which will place increased demand on the grid.

"Data centres can be powered by renewable energy and batteries. There is no technical reason why this isn't feasible. The unfeasibility will come from the politics," Srivastav said.

"The UK needs to build out more wind, more solar and more batteries ... and streamline the permitting and regulatory process."

What role could nuclear power play in the UK's AI push?

The government plans to test nuclear fusion as a way to power its AI ambitions, with the first AI Growth Zone at Culham, the headquarters of the UK Atomic Energy Authority (UKAEA).

The Tony Blair Institute for Global Change said in December that expanding nuclear power could generate 890 gigawatts electric (GWe) globally by 2050, compared with 369 GWe today.

Starmer's plan calls for the creation of a dedicated AI Energy Council to work with energy companies to understand the energy demands and challenges of AI development.

A government statement said this would support its mission to become a clean energy superpower by tapping into technologies like small modular reactors (SMRs).

SMRs use nuclear fission, have a lower capacity than do traditional nuclear power reactors and can be factory-assembled and transported.

The government plans to deploy "a first-of-a-kind" SMR by the early 2030s.

Around the world, large technology companies are looking at nuclear energy as a possible new power source for the vast storage units that fuel its products, especially AI.

Microsoft plans to restart the Three Mile Island nuclear reactor in Pennsylvania to help power its expanding data centres, while Google signed the world's first corporate agreement to buy power from multiple SMRs in October.

However, nuclear reactors are expensive, and the waste remains radioactive for hundreds of thousands of years. 

"(Nuclear power) will take a long time, and by that time it may be too late," Ramchurn said.

"The UK has to focus on the near term and look beyond its shores," he said, arguing that the UK should build data centres abroad in order to better access low-cost energy.

(Reporting by Adam Smith; Editing by Ellen Wulfhorst.)