Risky business: The agency helping finance the ideas of the future

Kathleen Fisher has spent her career at the bleeding edge of American innovation. Now, as the chief executive of the UK Government’s Advanced Research & Invention Agency (Aria), she is tasked with changing the way Britain approaches science funding. 

“Aria’s mission is to create breakthrough science to make the world and the UK a better place,” Fisher says during a visit to the National Robotarium, based at Heriot Watt University in Edinburgh. “We do that through programme-based science that is driven by programme directors. Every programme has baked into it both a ‘science miracle’ to disrupt science in some way and then a ‘cross the valley of death miracle’ because you have to not just do the science, but then you have to get it so that it actually delivers to the taxpayer and makes the world a better place.”  

Part of Aria’s mission, when it was announced by Kwasi Kwarteng, then the UK Government business secretary in 2021, was an acceptance of “higher tolerance for failure than is normal” to give innovators the confidence to pursue ambitious ideas, guided by the belief that “the freedom to fail is often also the freedom to succeed”. And with an initial budget of £800m at launch in 2023, rising to at least £1bn between 2025 and 2029, the agency was given enough financial firepower to take those risks in a market increasingly shaped by global venture capital firms. 

This appetite for financial risk was initiated by Dominic Cummings, Boris Johnson’s scandal-plagued former chief of staff, who pushed for the creation of an agency that could fly in the face of the usual Westminster approach to investment. Cummings modelled Aria on the US government’s Defense Advanced Research Projects Agency (Darpa), which exists to fund high-risk, high-reward research projects that often extend beyond military needs. 

So it’s no surprise that when Aria went searching for a new chief executive at the start of 2026, it turned to the agency that inspired it in the first place. Fisher, a former programme manager and director of the Information Innovation Office at Darpa, took on the top job in February after a career that included teaching posts at Tufts University and Stanford. 

In launching Aria, Cummings envisioned an agency that, like Darpa, could exist at a distance from the Westminster “blob”, independent of unnecessary bureaucracy and free to operate with the flexibility and speed required to compete on a global stage.

“One of the things that is important to keep in mind is that the science breakthroughs in the first place are really hard,” Fisher says. “Of course, things are going to fail all the time, and that’s fine. That’s a normal part of the process. It’s just some of the times we’re going to succeed, and those successes kind of pay for all the times we failed along the road. And I think the part that is important is to accept that that’s fine, right? And to not kick ourselves too much when it doesn’t work because what’s important is we keep going, we learn from the places we failed, we learn how to do it better and we don’t give up, we just keep going.”

Both Darpa and its predecessor, the Advanced Research Projects Agency (Arpa), were instrumental in creating technologies such as the internet, GPS and pioneering early work and investment in the development of vaccines used in Covid, which was directly related to a failed project exploring the possibility of synthetic blood.

“The programme manager sat back and looked at the technology that was created in the synthetic blood programme, and thought could it be used for something else,” says Fisher. “He invested in continuing to explore that technology by giving a small grant to a company in Massachusetts called Moderna. That technology was the mRNA vaccine. That was a failure from the perspective of the programme, but it was a massive success in terms of getting the world a technology that turned out to be really important shortly after.”

Kathleen Fisher of Aria | Image: Aria

In 2025, as part of this strategy of funding fringe technologies and research, Aria invested £1m into the robotics firm Touchlab, based at the National Robotarium. “We are making tactile sensors to give robots the sense of touch,” says Dr Vladimir Ivan, vice president of robotics at Touchlab. “We believe that this is one of the key missing components for robots to be skilful, dexterous and useful. All other animals and creatures out there, the main thing they have is the sense of touch, even if they don’t have any other senses, and robots don’t have that.”

This investment, part of Aria’s £57m Robot Dexterity Programme, is intended to create vastly more capable machines for use in a range of contexts. Touchlab aims to do this by retrofitting robot hands with an “e-skin” covered in sensors that can give localised readouts to the robot’s operating system.

“We’ve been building sensors, but our ambition is to build robots that can help solve actual real-world challenges,” Ivan says. “This is at the core of our company, and I think this is what really resonated with Aria, that yes, we do have a bigger plan and we’re not in it just for the money. And I think this is exactly what Aria is all about. It’s not just to fund ideas, but to help create solutions and not just in the UK, but globally. And that’s our ambition as well.”

An important aspect of Darpa’s success, and therefore something that Cummings looked to imitate in the establishment of Aria, is its fiercely independent position within the structure of the US government. “One of the things that Darpa talks about is how it’s really, really important to be off of roadmaps because it’s really important for the agency to establish its own priorities,” Fisher says. “For example, Darpa funded the research on unmanned aerial vehicles. The Air Force hated this idea and wanted to squash it entirely. But Darpa said ‘you can’t tell us to do that because we’re off your roadmap’. So that independence is really, really important.”  

To achieve this independence, the UK Government agreed a structure that gave Aria strategic autonomy over funding choices. This was outlined in an agreement signed by all four nations that said “Aria requires a minimal level of accountability for the spending of public money”, with additional operational autonomy over recruitment and a firm commitment that the body “will not be subject to the layers of approval and review that might exist elsewhere in the funding system or in government”.

This agreement also included an exemption for Aria from Freedom of Information (FOI) requests along with an overarching requirement that the “involvement of all four administrations of the UK be minimised”.  

This arm’s-length approach has led to criticism of the agency, with the News Media Association and the Campaign for Freedom of Information warning that excluding Aria from FOI requests will lead to the agency being viewed with “public suspicion and
distrust.”

“Aria was designed from the outset to operate independently in its programmatic research and funding decisions,” Fisher says. “As a public body, our independence comes with clear accountability to parliament, which in September included giving evidence to the House of Lords Science and Technology Committee. We’re subject to scrutiny by the National Audit Office and are required to submit an annual report and statement of accounts. Our board of directors also plays an important role in scrutinising our strategy and ensuring good governance. We do our work in the open, publishing every programme thesis and awarded project by default, so anyone can see who and what we fund.”

Part of the distrust around Aria comes from its explicit objective to pursue “high-risk, long-term” research and development that requires “significant risk appetite” with past projects including climate engineering, brain–computer interfaces, genetically modified crops, robotics and artificial intelligence.

The climate engineering programme, an investment of £50m into a programme of small-scale outdoor geoengineering experiments to generate data on the feasibility and risks of solar radiation modification (SRM) technologies, faced particular pushback. SRM technology works by reflecting a portion of incoming sunlight through the injection of reflective particles into the atmosphere or by increasing the reflectivity of clouds to reduce the effect of global warming. 

In written evidence presented to the UK Parliament by the Action Group on Erosion Technology and Concentration, the group warned that the research potentially puts the UK Government in contravention of the international moratorium on solar geoengineering research agreed through the UN Convention on Biological Diversity in 2010. The action group also stated that “Aria’s explicit ‘high risk-high reward’ approach flies directly in the face of the accepted use of the precautionary principle”, which they claim is intended to protect against publicly funded research resulting in unintended environmental and social consequences.

“In terms of the risk of unintended consequences, that’s always a risk when you’re pushing the frontier of technology,” Fisher says. “But we have a robust ethics and social responsibility part of the agency and that is part of the process when starting a programme. Every time a programme gets started at Aria, the programme director and the agency goes through a process of thinking about ‘what are the ethical and societal implications of this, and are they sufficiently high that we should set up a special process to monitor as the program goes along?’”

Just over three years into its existence, Aria remains an outlier in the way it approaches funding and research, pushing the boundaries at speed with a tolerance for risk and failure that is unique among government agencies. For Fisher, the need for such an agency is only becoming more crucial as technologies such as AI accelerate the pace of discovery. 

“I do think the pace of science will go faster because AI will help us to go faster,” Fisher says.

“Automation will help us to go faster but we do have to be careful as we do that because AI isn’t trustworthy. It’s like you have a very capable intern who is like a genius but also an idiot kind of at the same time. And you have to be very careful about how you supervise the idiot to take advantage of the benefits of the genius part of the intern.”

At the end of 2029, as part of Aria’s core commitment to term limits, Fisher's time as Aria's chief executive will come to an end if she doesn't apply for a new term. The time limit acts as an incentive to force programme directors to cut projects that aren’t working and double down on promising leads in a world where change comes quickly and often without warning. By that time, a lot of money will have been spent, some of it on projects that almost certainly will have failed. But if Fisher and Aria are right, somewhere among those failures could be the start of the next piece of breakthrough technology that could change the world.