Taking ‘bold’ bets: new UK agency prepares to fund breakthrough technologies

A new UK government agency set up to transform science by supporting high-risk, high-reward research is emerging from stealth mode with the aim of achieving a “world changing impact”, according to its chief executive.

Ilan Gur outlined ambitious plans for the £800mn UK Advanced Research and Invention Agency (Aria), which is seeking to invest in breakthrough technologies in fields from artificial intelligence and computing to neuroscience and materials science.

In the first interview since his appointment to Aria, Gur told the Financial Times that much of the agency’s success or failure would depend on its eight newly appointed programme directors who will select and fund projects. They will each have a budget of about £50mn.

Aria expects to unveil the identities of the directors next month and Gur said: “They are empowered to take bold, not safe bets. We have an opportunity here to create something that could have a world-changing impact for future generations.”

The establishment of Aria has been inspired by organisations in the US: the Defense Advanced Research Projects Agency (Darpa) and the Advanced Research Projects Agency-Energy (Arpa-E).

Darpa’s inventions include the key technologies behind the internet, satellite navigation and voice recognition. The achievements of Arpa-E range from sophisticated batteries to materials for nuclear fusion reactors.

These US agencies have programme managers taking decisions on the selection and funding of projects similar to the roles envisaged at Aria — there are no jobs at existing UK public research bodies with this level of freedom and flexibility.

“I was overwhelmed by the quality of the people who responded to our call for candidates,” said Gur, who joined Aria one year ago after a career in the US with start-up businesses and a spell at Arpa-E. “We have an incredible group who come from very different disciplines.”

The projects for which Aria provides funding will range across science, technology and engineering. They will involve scientists and engineers right across the UK research landscape — universities and public laboratories, start-ups and large companies — and some are expected to include collaborative funding with other agencies.

Gur and his colleagues at Aria are debating how they should tackle AI. One idea is to find an alternative computing technology to carry out AI, instead of the silicon semiconductors used exclusively today at a high cost in terms of energy consumption and climate change impact.

“Silicon isn’t the only way to do computation,” Gur said. “It happens in nature tens of thousands of times more efficiently, for example in the brain.”

The Aria team is considering whether an entirely new way of computing could be created specifically for AI rather than for general purpose computing. “We are exploring this area but we don’t know exactly what programme will come out of it,” he added.

Although Dominic Cummings, who served as chief adviser to former prime minister Boris Johnson, is widely regarded as the person who pushed the government to create Aria, Gur knew little about him before he took up the job.

“Since I arrived here [Cummings] has come up a lot,” he said. “What is great for me is that just about everyone I’ve talked to across different parts of the research community and across the political spectrum seems not only to be excited about Aria but to understand why it is important and why it is structured the way it is.” Cummings has not been in touch with him, Gur added.

The government and Aria’s founding team have faced criticism for not getting the agency going more quickly.

Plans to create Aria were announced in 2019, the initial four-year budget of £800mn was outlined in 2020, legislation to establish the agency received royal assent in 2022, and Gur was appointed last year. Then Aria was formally established as an independent research body in January.

But Gur said taking time to get Aria right was essential. The US model could not be transplanted straight to the UK without modification and the correct foundations had to be built — not only the programme directors and other core staff but also Aria’s governance structure.

The nine person board is chaired by Matt Clifford, chief executive of Entrepreneur First, a business which helps people to found new companies. Aria’s directors also include the venture capitalist Kate Bingham, former head of Britain’s vaccines task force, and Patrick Vallance, the former UK chief scientific adviser. Aria is based at the Alan Turing Institute, the national centre for AI research, in the British Library.

Aria’s initial budget of £200mn a year seems small both as a proportion of government research spending, which is due to reach £20bn annually by 2024, and in comparison with its US counterparts. Arpa-E received $470mn this year for advanced energy research, while Darpa spends $4bn annually on breakthrough projects for national security.

But Gur was confident that “as a catalyst” Aria will have enough money to work wonders.

“Our programmes will magnify and amplify the UK’s excellent existing science and engineering assets as a sort of force multiplier,” he said. “Aria is meant to operate on a decadal timescale and our budget allocation for four or five is enough to get started and achieve some early proof points that Aria is working.”

Gur did not reprise the government’s ambitions for the UK to be a “science superpower” but he recognised that part of Aria’s mission was to benefit the UK.

“I have an eight-year-old and a 10-year-old,” he said. “Success for me will be that, when they grow up, they should be able to look around the UK and see something really powerful — perhaps a new industry that has taken root — that was enabled by technology that Aria catalysed.”