The UK is promising a massive increase in computing capacity to build an AI industry

Stay informed about free updates

Just register to go Artificial intelligence myFT Digest — delivered directly to your inbox.

The UK will invest in a major expansion of the government’s AI computing capacity over five years including building a new supercomputer as it seeks to create a globally competitive artificial intelligence sector, ministers will announce on Monday .

This step is a response to a report recently published by AI UK economic opportunities, commissioned by the government and written by British investors Matt Clifford.

The supercomputer will join two other advanced UK machines including Isambard-AI at the University of Bristol, which has around 5,000 graphics processing units (GPUs), special chips to build AI software, with Dawn at Cambridge University.

Clifford’s report advocates reaching as many as 100,000 GPUs at the federal level by 2030.

The new capacity, which could represent a 20-fold increase in UK computing power, will be allocated to private AI data centers and will be commissioned by the government mainly for AI applications in education and services. the community.

It is not clear how much the project will cost, although it will come from the Department of Science, Innovation and Technology’s research and development funding.

The announcement comes as Clifford was appointed to the interim role advisor to ministers with AI, to help implement the recommendations in his report, according to two people briefed on the project. Downing Street declined to comment on the proposals.

Prime Minister Sir Keir Starmer said: “Our strategy will make Britain a world leader (in AI). It will give the industry the foundation it needs . . . That means more jobs and investment in the UK, money more in people’s pockets, and to change public services. This is the change that this government is bringing.”

Starmer raved about the value of AI as an engine of economic growth and social sector development after dinner with former Google boss Eric Schmidt and DeepMind head Sir Demis Hassabis on the UK’s first night . world financial summit in October, according to people briefed on the matter.

Clifford’s report, known as the “AI Possibilities Action Plan”, was submitted to the government in September, but its publication has been plagued by delays. Several cabinet ministers met to discuss its issues in December, according to people briefed on the matter.

It presents 50 recommendations to create a successful national AI industry by improving the conditions for building, measuring and adopting new technologies.

Among the recommendations accepted by the government are: creating AI “growth zones”, areas of the UK with faster access to planning permission to build AI infrastructure; and the AI ​​Energy Council, to advise on energy resource requirements for AI, including nuclear power.

Technologists, including Clifford, have argued that independent computing power is essential to ensure that British AI companies and researchers can become independent of AI businesses in other countries.

They argue that the potential could create new AI technologies and companies that matter globally, and that having access to reliable computing power at low cost is essential as computing becomes the arena of political struggle.

Science and Technology Secretary Peter Kyle came under fire in August for rejecting funding for the £800mn Exascale supercomputer program at Edinburgh university, a machine that can perform complex scientific calculations such as physics simulations, a move which he holds the technology and the faculty. the guard.

Kyle insisted he “didn’t cut anything”, as the £800mn promised by the previous government had not been budgeted.

Due to the lack of new independent software, the UK’s most powerful computer has been overtaken by rivals, meaning the country no longer has a top-tier machine. Worldwide 50.

Additional reporting by George Hammond in San Francisco