Unlock Editor's Digest at no cost
FT editor Roula Khalaf selects her favourite stories on this weekly newsletter.
The UK government has insisted it is going to proceed to fund breakthrough technologies in artificial intelligence and supercomputing, after sparking a backlash by cutting £1.3 billion in support for technology projects.
The latest Labour government under Sir Keir Starmer is desperately attempting to contain the fallout from its decision earlier this month to withdraw support for projects comparable to the development of an £800 million supercomputer on the University of Edinburgh.
Experts warn that Britain is lagging behind its international rivals in supercomputing, with ministers saying investment was not adequately funded by the previous Conservative government.
The funding row is a primary test for Labour as the federal government balances tight restrictions on public spending with its promise to advertise science and technology to assist achieve ambitious economic growth targets.
“Supercomputing continues to be a top priority for the federal government – it's a priority for (Technology Minister) Peter Kyle and he clearly wants us to get it right,” said one government official. “We want every bit of presidency money to make an impact. We should not going to present up on supercomputing.”
Exascale supercomputing is the power to perform one billion billion operations per second, often known as an “exaflop.” It is taken into account the holy grail of computing and is critical to the widespread adoption of AI.
There are two such fully functional exascale computers on the earth, each within the USA. Experts consider that China also has at the least one, even though it has not included it within the international rankings for computing capability. Major economies – from Japan to Europe to Saudi Arabia – are investing billions in similar projects.
Kyle told the Financial Times he was “preparing for a daring approach” based on an “AI motion plan” currently being developed by entrepreneur and AI expert Matt Clifford and attributable to be published in September.
He said the plan would “lay out our future computing needs and the way we are going to meet them.”
Clifford said last 12 months before he took a job for the federal government that “the UK doesn’t have very large sovereign computing capability” and that it should “spend billions over the subsequent decade to be certain that crucial AI firms are in-built the UK”.
The UK has already invested £225 million in an AI-focused supercomputer being developed on the University of Bristol. It known as Isambard-AI and is ready to grow to be one of the crucial powerful exascale computers in Europe. Two other supercomputers – Archer2 on the University of Edinburgh and DiRAC, built at several universities – will reach the top of their life in two years.
Two senior government insiders told the FT that Starmer's ambitions for Britain's supercomputer potential have certainly not been dampened, but there are some concerns about how the present budget will probably be distributed.
A senior government official noted that Edinburgh's exascale project made little strategic sense since it was focused on more traditional computing projects comparable to scientific simulations reasonably than AI. The person added that the brand new government was expected to fulfill previous spending commitments on supercomputing within the upcoming autumn budget.
But shadow science minister Andrew Griffith told the FT it was a “misrepresentation” by Starmer's government that the project was “unfunded” because the investment would all the time be included within the autumn 2024 budget. “It's an enormous setback for British technology and… it appears to be an indication of waning ambition with regards to Britain's plans to grow to be a science and technology superpower,” he said.
Several supercomputing experts the FT spoke to agreed that investing almost exclusively in a single sort of supercomputing hardware was not the “most strategic” move, and that the brand new government should as an alternative focus its resources on a variety of supercomputing hardware, software and skills to fulfill Britain's AI and science needs.
Mark Wilkinson, director of the DiRAC facility and a theoretical astrophysicist on the University of Leicester, said it was price pausing and serious about the very best option to invest limited resources, on condition that “the UK has fallen far behind comparable countries by way of computing capability”.
He warned that the federal government shouldn’t “slice up the funds in a salami-like manner, so that everybody only gets a small amount”, but as an alternative deal with the very best projects. The deal with “exaflops” is misplaced within the UK, he said, adding: “The variety of flops is only one measure of performance – what ultimately matters is productivity.”
Sir Peter Mathieson, chancellor of Edinburgh University – which spent £31 million of its own money to construct its exascale supercomputer – has to this point kept away from publicly criticising the federal government's decision. He met Kyle and said he welcomed “the continued dialogue on this essential issue”.
Nevertheless, several circles strongly criticized the way in which by which the Starmer government communicated its decision to chop the funds without promising a substitute.
Ben Johnson, a former policy adviser to the last Tory technology minister, said “unclear communication … may lead to confusion within the research community.”
“The market is amazingly hot at once and is getting hotter by the week,” he added. “Any delay in procurement could possibly be very costly.”
