In a cramped conference room in Bristol, Ilan Gur is trying to convince a group of plant biologists that they can change the world. The 44-year-old has the patter you’d expect from a Californian startup founder, but he’s also one of the UK’s most senior civil servants, so what comes next is unexpected.
Close your eyes, he asks the scientists, and imagine pushing past the very edges of your research. The attendees take a beat, shifting slightly on their uncomfortable chairs. Positive visualization is not quite what they had expected from a workshop introducing them to the Advanced Research and Invention Agency (ARIA), the UK government’s new high-risk, high-reward science funding agency.
Gur is ARIA’s CEO, and if he senses their hesitation, he is unfazed by it. The whole point of ARIA is to push researchers beyond their comfort zones and towards ideas the typically risk-averse British science funding system would deem improbable or downright weird. Today, it’s plants with genomes written from scratch to grow foods, materials, and medicines that don’t yet exist. Tomorrow it could be ways to cool down the planet or build more dexterous robot bodies. The plan should be just on the edge of impossible, Gur tells the room. Impactful enough that it’s worth a shot, but so ambitious that half of the scientists leave the workshop convinced it’ll never work.
ARIA is designed to put Britain back on the scientific map. By the mid 2010s, the birthplace of Isaac Newton and Alexander Fleming had become, in the views of some insiders, sclerotic and backwards looking. Inside Downing Street, government advisers were looking toward the US and wondering why the UK seemed such a laggard when it came to truly transformational scientific breakthroughs: Crispr gene-editing, mRNA vaccines, most major AI research (with the notable exception of DeepMind), all happening outside the UK.
Inspired by ARPA, the US agency that helped birth what became the internet, GPS, and the era of personal computing, ARIA is an attempt to find a new way of funding breakthrough British science. It’s designed to be ambitious and nimble, and to take big risks. Its employees have an extraordinary amount of freedom over how and who it will fund: startups, universities, individuals, anything is on the table. Its senior employees are exempt from the ordinary restrictions on civil service pay. (Gur’s salary of between £380,000 and £385,000 ($510,000) makes him one of the highest-paid civil servants in the country.) Other agencies send out departmental press releases. ARIA fires out Substack updates.
“We are looking for things that are controversial and risky in terms of whether or not they might work,” says Angie Burnett, a plant scientist who joined ARIA in October 2023 to lead the agency’s work on synthetic plants. Burnett is one of eight program directors with tens of millions of pounds to spend to fund breakthroughs in their own scientific niche. Her colleagues are searching for new ways to safeguard against dangerous AI, measure climate tipping points, and manipulate the human brain. If any one of these bets comes off—and ARIA staff agree that many will fail—then the benefits should massively outweigh the £800 million ($1 billion) of public money the agency has been allotted for its first four years.
Burnett joined ARIA from the University of Cambridge, where she studied how crops adapt to stressful environments. “I left an organization that was over 800 years old to join an organization that was months old at the time, and that is a huge difference,” she says. In her application she pitched insect-size drones that could monitor individual plants—a doctor for every plant, she imagined. That idea was a nonstarter, but she eventually settled on the challenge of developing synthetic plants with genomes written by human hands. Soon she will start funding the scientists trying to create fully synthetic chromosomes—a breakthrough that, if successful, would be a leap beyond the current capabilities of plant scientists.
“I’ve always had this desire for impact,” says Burnett, who is 34 and whose love of greenery extends to her wardrobe—she’s wearing a sleeveless top patterned with plants. She talks with the earnest, unfiltered enthusiasm of someone who is genuinely excited about her work, even if its ultimate destination is still somewhat uncertain. Before Cambridge, Burnett worked for the Food and Agriculture Organization of the United Nations in Rome, where she focused on helping smallholder farmers in Asia and Sub-Saharan Africa. But it is at ARIA that Burnett thinks she could end up having the biggest impact of her career. “I think the potential is enormous,” she says.
That is, if anyone can agree on what synthetic plants are actually for. In Bristol, Burnett has split the scientists into smaller groups to discuss her program. One group is throwing out possible ideas for what one might do with a fully synthetic plant. Bus shelters made from crops? Too gimmicky. Edible air-conditioning systems? Weird, but intriguing. At the head of the circle a civil servant takes notes while the scientists careen from one improbable idea to another. Dolphins with lasers … but plants?
The seed for ARIA was planted during one of the most chaotic periods in modern British history. In mid-2019, the UK was crashing out of the European Union and Boris Johnson was crashing into 10 Downing Street. On Sunday July 21, Johnson—the former London mayor who was the favorite to win the Conservative leadership election and become the next prime minister—entered a home on a quiet north London street to cut a deal. The UK’s departure from the EU had already brought down two Conservative prime ministers, and Johnson was determined not to be the third. To make sure of that, he needed the help of Dominic Cummings.
Cummings was an acerbic political adviser who had spearheaded the 2016 Vote Leave campaign during the EU referendum. He was also a prolific blogger, publishing long essays that railed against what he saw as a turgid, mostly incompetent civil service that he now calls the “deep state.” But he was also fascinated by organizations that transformed the world: the Manhattan Project, Xerox PARC, Bell Labs, ARPA and early NASA. Mid-century America seemed to buzz with breakthroughs. But since then—in Cummings’ estimation—science funding had become more cautious, more bogged down in bureaucracy, and less creative, particularly in the UK. (Cummings did not respond to WIRED’s interview requests.) In one 237-page-long screed from 2013 he called for a high-risk/high-reward agency focused on energy projects that “operates outside all normal bureaucratic systems,” In the following years he returned to the topic multiple times. If the UK government only cared about science and progress, this inertia could be reversed, Cummings hoped.
Amid the Brexit chaos he’d helped deliver, Cummings saw an opportunity to make that happen. The crux of the deal that Cummings presented to Johnson was this: He would become the prime minister’s chief adviser and help stop Brexit subsuming Johnson’s premiership, and in return Johnson would double the science budget, avoid a second referendum, overhaul Whitehall, and create a breakthrough funding agency modeled on ARPA. Cummings’ reported WhatsApp status during this time hints at his priorities: “Get Brexit done. Then ARPA.”
Curiosity about a better way of funding science extended well beyond the Cummings blogosphere. “There was quite an appetite for thinking around different ways of funding,” says Patrick Vallance, who was the UK’s chief scientific adviser between April 2018 and 2023, and one of the most visible faces of the government’s pandemic response. Vallance would later become a founding member of the ARIA board, before leaving in July 2024 when he was appointed by the new Labour government to be minister of state for science, research, and innovation (he spoke to WIRED before leaving ARIA). By the time Johnson came into government, Vallance says, a lot of people were starting to think hard about how the new funding agency, which still lacked a name, would work. “UK ARPA was the name they called it, which was driving me nuts,” Vallance says. To him, UK ARPA sounded pathetic—the point wasn’t to replicate ARPA, but to do something new.
One of the key disagreements was about what problem the new agency was supposed to solve. Most government-backed science is funded through UK Research and Innovation (UKRI), which was set up in 2018 to simplify a byzantine funding system of seven research councils and numerous other bodies. But UKRI already oversaw the so-called Catapult projects that were supposed to fund ambitious, transformational research that other agencies might miss. Some thought that the new agency should focus on translating scientific breakthroughs into real-world progress says Vallance, while others thought the focus should be entirely on “blue-sky” research.
On January 2, 2020, Dominic Cummings published a blog calling for “weirdos and misfits” to come and work with him in 10 Downing Street. “I’ll bin you within weeks if you don’t fit—don’t complain later because I made it clear now,” Cummings warned. One of those who responded to his call was James Phillips, then a neuroscience researcher who had met Cummings four years earlier, after sending a response to Cummings’ blog-screeds on the state of UK science, and again in 2018 at the Janelia Research Campus in Virginia. “I walked into Number 10 and that was my first-ever office job. It was my first job in politics, it was my first job in policy,” Phillips says. He joined in April 2020, just as the UK grappled with the first Covid wave. About a week before Phillips entered Number 10, Boris Johnson had been in intensive care with a near-fatal case of Covid-19.
Phillips avoided being binned by Cummings, but his days were consumed by the government’s response to the coronavirus. “Thinking about ARIA was part of the day when it wasn’t dealing with the horrible situation that we faced,” he says. One of the key questions was whether it would be a so-called “mission-driven” agency, with a clear direction from the government to find breakthroughs in specific areas, like medicine or clean energy. A report by the House of Commons Science, Innovation, and Technology Committee published in February 2021 argued that ARIA’s focus should be determined by the government, not the agency.
Phillips and Cummings balked at this idea. “We felt strongly that the government should not be telling ARIA what to focus on,” Phillips says. Even though ARPA eventually limited its remit to defense projects when it first became Darpa in 1972, the agency still took a broad enough approach that in 2013 it awarded Moderna up to $25 million to research and develop mRNA vaccines. The team setting up ARIA feared that a narrowly defined agency would end up missing these tangential projects that turn into unprecedented world changers—exactly the kind of research that they wanted to encourage.
The team at Number 10 also wanted to avoid the mistakes of other countries that had tried their own breakthrough agencies. “All the countries that have tried to replicate an ARPA-like model have kept it on too short a leash, and that kills it,” says Vallance. “The tendency in government very often is to get the leash shorter and shorter and shorter […] because it’s public money, and people feel a responsibility, and it was new and shiny and everyone wanted to be involved.”
The bill setting up ARIA—The Advanced Research and Invention Agency Act—became law in February 2022. The document is the closest thing the agency has to a sacred text. It’s at the top of the reading list Gur provides to new board members. When I meet ARIA chief financial officer and COO Antonia Jenkinson, she is carrying an annotated copy of the act. It’s all in there, she says. It’s independent, with future CEOs appointed by a board’s chair, not the government. It’s shielded from a formal government review for at least 10 years, giving it room to pursue projects that might not pay off in the short term. It’s exempt from Freedom of Information laws that require other government agencies to provide public access to information. The point of all of this, Phillips says, is to keep ARIA as small and nonbureaucratic as possible.
It’s a remit that—combined with its sheer newness—makes it difficult to put ARIA in a box. ARIA staff often refer to the agency as a startup, albeit a startup funded by British taxpayers and staffed by civil servants. “We were a startup, we had to get the culture into the organization,” says Jenkinson, who joined ARIA in January 2023. “Every system, process, and person that I hire is all about trying to get the right appetite and the right culture into the organization.”
When I visited ARIA in spring, the agency was squeezed into a single room on the fourth floor of the British Library, an imposing red brick building a short walk from London’s King’s Cross station. Staff perched on the heavy windowsills as they listened to a colleague run through a series of milestones for the fledgling organization. Each one was greeted with scattered applause. New workshops! Clap! New trips planned! Clap! New joiners! Clap!
The team had expanded by a quarter in the previous six months, and it was clear the agency was starting to outgrow its single room, on lease from the nearby Alan Turing Institute. On some desks in the hallways, sticky notes reminded staff that they were not for ARIA employees. Securing a meeting room had become something of a dark art. Although ARIA had officially been formed in January 2023, the eight program directors at the heart of the agency had only joined that fall, and the agency had the feel of an organization still finding its cultural feet.
A lot of ARIA’s success—cultural and practical—will come down to the program directors, who largely decide who they will fund as they try to coax their vision of the future into reality. Each one will have around £50 million ($66 million) they can give to startups, individuals, or university researchers, among others. They can set up prizes, give out small seed grants, or fund much larger projects. Suraj Bramhavar, who joined ARIA after cofounding a cloud computing firm spun out from MIT, is running a program that aims to make AI hardware a thousandth of its current cost. Even pulling off a 10th of that headline number would “change the world,” says Matt Clifford, a British entrepreneur who chairs the ARIA board. “The test for any one program, is if only this program succeeded, if everything else failed and this was the only thing we could point out for the first decade of ARIA’s life, would it be worth it?”
Before Jacques Carolan joined ARIA, he was on the road to becoming a professor of physics. “That was my dream,” he says. But at MIT he started to talk to people—often former academics—who ran Darpa programs, and who had a certain lore and aura. When he heard that the UK was setting up its own breakthrough science agency, he knew he had to apply.
Now Carolan is the director of a £69 million program to develop more precise ways for us to interact with the human brain—perhaps smarter brain implants or stem-cell-based systems that could help treat disorders like treatment-resistant depression. One idea is to come up with therapies that are less invasive than the deep brain implants used to treat Parkinson’s disease and epilepsy, but are still precise enough to have a profound impact on the brain. “When you think about the potential impact you can have, there are not many times in your life where you have serious resources to realize a scientific vision of your own making. To realize your most ambitious vision,” Carolan says.
Jenny Read used to study praying mantis vision at Newcastle University—one time fitting tiny 3D glasses to the insects to figure out how their vision differed from our own—and whether understanding insect vision could lead us toward more efficient ways of building robotic vision. “I did that for many years and really enjoyed it, but I think I sort of reached a point in my life where you turn 50, kids grow up and leave home,” she says. Read was wondering what to do with the rest of her career when she decided to apply to ARIA. “I kind of felt a hunger in myself for being part of something bigger.”
Every other week the program directors meet in ARIA’s office. On the day I visit, they kick off a meeting with a little game. They are each given a superpower—invisibility, superhuman strength, time manipulation, and so on—and have to convince the others why their power would reign supreme. One program director is assigned telekinesis, which they imagine using to manipulate football games, winning vast sums of money, and increasing ARIA’s budget. (A little later the program directors are contemplating ways their programs might fail when they break for another quick game—a riff on musical chairs. This is an organization that takes ice-breakers extremely seriously).
A little telekinesis may come in handy. ARIA’s £800 million budget is tiny compared to the £25.1 billion budgeted for UKRI between 2022 and 2025. In 2024 alone, Darpa was allotted around $4 billion, and yet ARIA’s remit is—in theory at least—much wider than the US agency's. It is a modest budget for a big mission, which, according to Clifford, isn’t just to fund breakthroughs but to help elevate the status of science in the UK. “I think we’ve done a very poor job of elevating the status of discovery … And yet when you look at almost everything we value, it’s downstream of scientific progress. Without scientific progress you don’t get it.”
Ilan Gur has spent his career figuring out how to make that progress happen. In 2005, he was partway through his physics PhD in Berkeley when he decided he was more interested in the gritty work of turning breakthroughs into impact than a career as a scientist himself. Gur had been working on a way to use nanocrystals to make very cheap solar panels, and one of his papers had just been featured in the prestigious journal Science. He thought these ultra-cheap nanocrystals might change the world—if people could print solar panels the way they did newspapers, abundant green energy could be just around the corner.
And yet Gur realized that his nanocrystals stood no real chance of having an impact. They might be able to create dirt-cheap solar cells, sure, but the crystals didn’t last long, and the cost of reinstalling them every three to five years would be astronomically more expensive than conventional solar panels. “We could give those solar cells away for free, and they would not compete with the cells that you could buy at the time,” he says. “That’s actually a pretty simple analysis to understand the value of the technology that you’re trying to build, which no one in my lab had ever even talked about.”
It wasn’t enough to tinker away in a lab and hope your work would change the world, Gur was realizing. “If what you care about is not simply a scientific breakthrough but a scientific breakthrough that can catalyze world-changing impact in a tangible way, then there is so much in that journey. The breakthrough is necessary but not sufficient.” He would eventually leave academia and work at ARPA-E, a US Department of Energy organization modeled after Darpa to pursue high-risk, high-reward research in energy technology. Later he set up Activate—an organization that supports early-stage scientists working on potentially groundbreaking research. And then in August 2022, Gur moved from California to take the role of ARIA CEO.
Gur was a natural fit for ARIA, says Stripe CEO Patrick Collison, who is one of the agency’s advisers. Collison had only met Gur a couple of times before becoming an ARIA adviser, but he knew that the CEO was no armchair innovator—Gur was focused on actually getting shit done. It’s one thing to engage in abstract philosophizing, it’s quite another thing to conjure up a functioning institution and make it work, says Collison.
But it’s also part of Gur’s job to help build the myth of ARIA—to pull in talented scientists from across the world, and achieve the agency’s secondary mission of positioning Britain as a scientific leader again. The nuts-and-bolts of progress can become obscured by the narrative-building that surrounds genuine breakthroughs: J. Robert Oppenheimer and his boys in Los Alamos got the Hollywood treatment, but the lion’s share of Manhattan project dollars went to the complex in Oak Ridge that enriched the uranium for that first atomic bomb. The myth of Darpa, in particular, looms large over ARIA. “Darpa has done brilliant things and changed the world in a lot of ways, but it’s done that in the past,” says Gur, who wears the startup-founder uniform of sneakers, comfortable trousers and, each time we met, an ARIA T-shirt.
There are signs—beyond the branded clothing—that ARIA is generating its own myth. In August, it announced that the deep-learning pioneer Yoshua Bengio would join ARIA’s program on safeguarded AI as scientific director, working with program director David “davidad” Dalrymple. In Gur’s estimation, it was the prospect of working with Dalrymple that pulled Bengio into ARIA’s orbit.
Perhaps the very best myths become their own engines of progress. Clifford has a six-year-old son who very much wants to become a racing car designer one day. What if future children aspired to be the kind of person who nudged a new breakthrough into existence? Not scientists in the white-coat-wearing way of the imagination, but conductors of progress—engineering scientific crescendos that otherwise may have petered out into silence. “I kind of want people to say that they want to be an ARIA program director the same way that they say they want to be a Premier League footballer,” Clifford says. That, in its own way, would be a profound kind of success.
In the Bristol workshop, the scientists are introducing themselves. It’s an eclectic bunch: One person is an amateur flower designer, another is an architect interested in using plants as building materials. Some work in startups, others for big corporations. In the course of coming up with her program, Burnett says, she has spoken to over 100 people in and around the world of plant science. Today a decent proportion of them are in the same room for the first time—and some of them are sizing up whether they might want to get in on Burnett’s synthetic plants plan. As much as Gur and Burnett are angling for ideas, they’re also pitching the very concept of ARIA itself.
One of the plant scientists in attendance is Jane Langdale, who is perhaps the closest thing the field already has to a moonshot chaser. Langdale leads the C4 Rice Project, one of the most ambitious plant science projects ever undertaken. Crops like maize use a photosynthesis pathway that makes them—especially in hot, dry environments—more efficient than rice. The C4 Rice Project aims to recreate this C4 photosynthesis pathway in rice, supercharging the growth of one of the world’s most important crops. Since 2008 the project has been funded by the Bill & Melinda Gates Foundation.
“It’s great that ARIA exists, and I think it’s great that there is a plants program in it,” says Langdale. “There’s no doubt about that because for far too long, people like Gates have been driving the moonshot projects, and of course they have a very specific focus on what it is they want to achieve.”
Philanthropic foundations like Gates’ also have a higher tolerance for projects that may not hit paydirt. “We’ve been going quite a long time, and we certainly don’t have anything near a product to put in the field,” Langdale says. Government-backed science funding has historically had much less of an appetite for these kinds of projects, because it’s hard to justify spending taxpayer money on projects that might take 30 years to come to fruition.
Even compared to the C4 Rice Project, Burnett’s synthetic plants program is a very significant chunk of money, Langdale says. Burnett is aiming to spend £62.4 million ($82 million) over five years. The program will fund scientists to try to make synthetic chromosomes, the genetic building blocks of plants, and synthetic chloroplasts, which have their own separate genomes. But the program doesn’t specify what new features these partly synthetic plants should have. It’s a little like designing a new machine without knowing what tools that machine is going to build, says Langdale.
Johnathan Napier, a science director at agricultural institute Rothamsted Research shares these concerns. Building synthetic chromosomes and chloroplasts are clearly defined goals, but he’s not sure whether they’re going to deliver a tangible benefit. Napier tries to engineer crops to produce omega-3 fish oils, while the C4 Rice Project is attempting to make rice much more productive. But Burnett’s program is much wider than either of these. In theory at least, it could one day allow plant scientists to plug in any kind of functionality into a plant.
“If this all worked, you’d be able to design your complex pathway in the computer, build an entire chromosome […] and just plug that into the plant in a single step,” says Saul Purton, another workshop attendee and a professor at University College London who works on synthetic chloroplasts in algae. Purton says that he may apply for an ARIA grant, but that the five-year timeline set out to deliver synthetic chloroplasts in several crop species is extremely tight. “We’ve been bashing away in terms of developing new synthetic biology tools for engineering the chloroplast of a simple model system for 15 or 20 years now, and we’re still learning, we’re still making mistakes.”
When I meet Burnett again in early August, she has just had her program approved after a grueling three-hour meeting with Gur, members of ARIA’s executive team, and a panel of external experts. “It was a little nerve-wracking because it’s such a big moment that I’ve been working towards for this whole time,” she says. As well as funding projects working to build synthetic chromosomes and chloroplasts, Burnett is also asking for research into the ethics of synthetic plants—anticipating a world where farmers, lawmakers, and the public may have to grapple with the idea of crops fully crafted by human hands. But it’s unlikely she’ll still be with the agency to see those scientific seeds bear fruit. Program directors are typically hired for three-year terms, and the agency is already hiring its next batch of directors, some of whom will launch entirely new project areas.
Over such short timescales, it can be difficult to gauge the success of such long-term plays: Are mistakes just bumps in the road, or signs that you’ve taken the wrong route altogether? Collison is wary about defining success at all. Give it 15 years, he says, and it should be pretty obvious if ARIA is a good thing or not. The agency has a little breathing room. It cannot be dissolved for at least 10 years, by which point the UK will have had at least one more general election. The new Labour government has indicated its support for ARIA, not least by making Vallance the minister responsible for ARIA. “It is essential to harness the power of science to deliver economic growth, opportunity, and scientific advancements for people across the UK,” said a government spokesperson.
But discerning the arc of progress takes time. In 1958 ARPA might have looked like a desperate attempt to wrest back technological supremacy from the Soviets. That was true in a way, but its ultimate impact reached far, far, beyond that. Is ARIA an attempt by a legacy-obsessed prime minister to reinstate the UK’s relevance in a post-Brexit world? Well, yes, but that may become just the preamble to the agency’s real story.
In 1994, Gordon Moore published an essay that he called “The Accidental Entrepreneur.” In it, he describes how he seemed destined for a midlevel managerial role until a company psychologist at the chemical giant Dow told him he would never manage anything. Instead, he ended up cofounding Fairchild Semiconductor and then Intel, whose silicon chips underpin the exponential growth in computing power that’s driven much of our technological development over the last half century. In the fullness of time, Moore’s rise seemed as inevitable as the law he lent his name to, but to him that was never the case.
Every breakthrough eventually boils down to brilliant people working on fiendishly tricky problems. For Gur, ARIA is about building a system to lift up people like Moore, rather than leaving them to drown in academic bureaucracy or fizzle out in the startup sector. There was something innate in Moore that allowed him to be who he was, Gur argues, but he needed that accidental moment to make him great. “So what do I hope ARIA can be? It can be this sort of catalyst intervention which takes someone who’s got the latent mindset and spirit to change the world—and it’s the accidental piece of their story. Maybe it’s not so accidental anymore.”
This article first appeared in the November/December 2024 edition of WIRED UK.
