🏦 Ian Ng, Researcher

For those of us who follow Europe’s startup ecosystem, it won’t come as news that the late-stage funding gap is often seen as the binding constraint on startups scaling up. Much of this has already been mapped out in the Draghi report and elsewhere — the internal barriers within the European Single Market and the need to funnel wealth towards risk-taking VCs who can back scale-ups. While the Draghi report stated that the EU’s productivity growth over the past two decades could be on par with the US if you strip out the tech sector, a new NBER paper finds that the valuation gap between US and European firms persists even when comparing companies within the same narrow industry. Strip out the very largest companies in each region and the gap barely budges, so this isn’t mega-caps skewing the average. Instead, the authors identify the gap as concentrated among young, small, R&D-intensive companies — what stockpickers would call growth stocks, whose value lies in future growth rather than today’s cash flow.

The paper attributes the gap to two constraints. On the demand side, European firms’ sales are tied to the size of their home country’s economy. A 1% rise in local GDP raises a European firm’s sales by 0.8%, a dependency that does not exist for American firms and their home states. On the finance side, Europe has less venture capital than the US. US venture capital investment ran at 0.5% of GDP in 2023, against 0.15% even in Denmark, the most VC-active country in Europe. The paper links this to a structural difference in the supply of long-term capital between the US and Europe, with US retirement assets mostly in defined-contribution (DC) schemes carrying heavy equity exposure, whereas European pension systems remain mostly pay-as-you-go.

In the UK, the state pension is pay-as-you-go but workplace pensions have shifted towards DC schemes. There is progress on turning that pot towards venture capital. The Mansion House Accord commits 17 of the UK’s largest DC providers to allocate 10% of default funds to private markets by 2030, half of it in the UK, while the Pension Schemes Act 2026 gives the Government a reserve power to require providers to hit the target if voluntary progress stalls.

The private sector has already made the shift from defined benefit to DC. Given the ballooning state pension bill is at around 5% of GDP and forecast to climb towards 8% of GDP by the early 2070s, perhaps it’s time to ask the same question of our state pension. Our state pension is simply moving cash from today’s workers to today’s pensioners without investing a penny of that money into British growth. Germany has kickstarted its pension reform in this direction, as the government-appointed pension commission has recommended a compulsory contribution starting at 0.5% of pre-tax income and rising to 2% by 2031, split evenly between employer and employee and paid into a centrally managed public fund modelled on Sweden’s.

Europe is not broke — we remain among the wealthiest continents — but we must put that wealth to work. Closing the gap needs businesses willing to take risks and more capital willing to back them. It is essential for the UK and Europe to plug our gap in late-stage funding if we want to keep our best scale-ups.

🏛️ Eamonn Ives, Associate Director, Public First

Whether it’s Alex Depledge advising the Chancellor of the Exchequer, Matt Clifford holding the pen for Britain’s AI Opportunities Action Plan or Emma Jones taking the reins as the Small Business Commissioner, entrepreneurs going into the heart of government seems to be in vogue. Most readers will probably regard this trend as an unalloyed positive. In addition to bringing valuable domain-specific expertise, founders tend to have an uncanny knack for challenging established ways of working and breaking through the institutional inertia that besets modern governments.

A new working paper from Aaron Chatterji, Jorge Guzman, Joyce Ma and Ryan C. McDevitt puts some numbers behind the theory. Studying a dataset of American state legislators, they conclude that those who own and actively manage a firm not only propose a larger proportion of bills as first or sole primary sponsor but also “selectively advance pro-entry legislation, especially bills related to deregulation and innovation.”

Entrepreneurs, it seems, really do bring a particular perspective on economic growth into politics: one focused less on protecting the interests of incumbent companies, and more concerned with making it easier for new businesses to emerge. Perhaps due to having experienced the frustrations of starting up and running a firm first-hand, they seem especially attuned to how rules, regulations and institutional sclerosis can stand in the way of new entrants.

American state politics is not Westminster, so we should be wary of assuming the paper’s findings travel intact. But they ought to make us think more seriously about entrepreneurial experience as a form of talent that government should actively seek out. Political parties could do more to encourage founders to stand for office; governments could make it easier for them to spend periods inside agencies and departments through appointments, fellowships or advisory roles; and policymakers could be more deliberate about bringing entrepreneurs into the room before rules are written, rather than consulting them as an afterthought.

📍 Sophya Mashkoor, Researcher

The internet was supposed to kill the postcode, so why hasn’t AI finished the job?

With access to a computer, an AI coding agent, a Stripe account and a Zoom link, every founder can easily build their business from anywhere in the world. A great engineer in Lagos could build the same thing as one in San Francisco, and perhaps at a cheaper cost. However, in practice the map has gotten lumpier.

Findings from earlier in the year show that around two-thirds of US venture dollars land in the Bay Area, which takes roughly 60% of global AI funding. New York, Los Angeles and Boston take most of the rest, and everywhere else is left to split the crumbs.

So why are the tools that were meant to make location irrelevant the same ones pulling capital back into one postcode (or zip code, if you will)?

The old case for hubs was to cluster where the engineers are so talent is concentrated in one area. Now, with AI, a two-person team anywhere can ship what used to take a department, so the need for density weakens. But when building takes just an afternoon, engineers stop being the scarce resource, and it instead becomes the judgement of people who can tell whether an idea is worthwhile or not. Those people — repeat founders, ex-lab researchers, investors on their third AI cycle — are still mostly having coffee in the same few square miles.

So the idea that AI democratises entrepreneurship and that capital keeps re-concentrating in a handful of cities can coexist. AI has shifted gatekeeping away from capital and infrastructure but ideas are still better communicated in person.

It is therefore worth asking: is this temporary, or does judgement being scarce mean hubs never really dissolve?

There’s a kind of founder network in the UK that meets after 5pm, usually in the pub, and is attended by people who have nobody waiting for them at home.

It’s a world Maria Kardakova once felt she had no choice but to live in, staying out late to keep pace with her peers, all while worrying about after-school pickups and whether she was spending enough time with her two young children.

For female founders, that expectation is a tax. Around 70% of UK venture capital deals still happen through informal networks like these.

Maria says lockdown gave her the excuse to stop, and she’s glad it did. Because in her newly free evenings, she realised the room didn’t have to be the one with everyone else in it.

With her two children at home and a husband recovering from cancer, she found herself overpaying for a meal-kit box and becoming frustrated by half-cooked recipes nobody actually wanted to eat. Out of that frustration came iCook, the AI nutrition platform she’s since built into a 250,000-user business across 176 countries, integrated with retailers from Tesco to Walmart.

She built it on a set of firm refusals, most of which she still holds. We spoke about what it costs to build outside a system designed for someone else, why she won’t sell supplements, and why she’s waiting for a Series A to say what she thinks.

What we discussed

• How a lockdown argument with a meal-kit box became a 250,000-user nutrition platform.

• Who can afford to network: 7.30pm events, self-funded pitch trips and the founders who aren’t in the room.

• Why every scientific nutrition platform pivoted into supplements and why she won’t.

Lessons for policymakers

• Childcare is startup policy. In Sweden, leave is genuinely shared, childcare from year two is good, and the state pays the same per child regardless of income.

• Nothing sits between retailers and what they promote. Over 70% of food advertising spend goes to confectionery, sugar, salt and sweet snacks; under 2% to fruit and vegetables.

Lessons for founders

• Hold your ground. She was told more than once, in terms she calls mansplaining, that recipes were finished and ChatGPT would replace them. She smiled and nodded, but now thinks she should have pushed back.

• Chase. Scientists are trained to let the work speak and wait to be approached. She now chases investors, and prepares a market proposal in hours where she once would have taken a month.

• Boldness has a price. She’s holding her opinions until Series A, on the theory that a funded founder with opinions is read differently from an unfunded one.

Lessons for event organisers

• Children’s rooms beat lactation rooms for a founder doing back-to-back meetings.

• A free ticket for a spouse or nanny changes who can attend.

• A founder invited to pitch is usually paying their own airfare.

The founding story

You have a doctorate in biosciences and a decade across food, retail, catering and biotech. Where did iCook start?

The idea came in 2020 when I was stuck at home with my two younger children and my husband, who had recently had cancer.

I had to feed the whole family, and we tried different boxes and within a couple of deliveries we realised we were overpaying by five times for peeled carrots or some garlic cloves. The food then gets stuck in the fridge, and nobody wants to eat it because it’s already half-cooked and prepped. You still end up spending the same amount of time and money on cooking and buying at the regular supermarket.

So I thought: why not generate a balanced menu instead? A simple machine-learning model that takes the family’s allergies, ethnic background, ages and size, builds the menu, works out how much of each ingredient you need and sends it to the shopping list.

What did the next two years look like?

Execution came in 2022, when I found an investor for £50,000. I already had an iOS prototype, which we launched as a paid app; revenue from that paid for the Android version, a designer, and a food photographer — we had our own kitchen for production and I wanted a minimalistic look. We ran ads on Amazon, the App Store and Google Play to work out which markets to expand into, and broke even within a year on £100,000 of revenue, which went back into the AI.

2024 was the groundbreaking year, when a helpful household tool turned into something closer to an agent. I built the infrastructure myself — probably too precisely — but it means I have hands-on knowledge of every ingredient, policy and nutrition recommendation behind it.

My biotech background means I also know GDPR, so I know what we can and can’t share with third parties, and what sits on Google Cloud versus AWS. After that we put money into the AI rather than advertising, brought a few more people on, and we’ve been raising ever since.

What are you building now?

We’re testing a digital avatar version of the AI assistant. It isn’t just a Duolingo for cooking, teaching you nutrition from a young age. It also gives advice if you’ve been diagnosed with something.

The avatar can support you instead of you going to social media for advice, which is why we want to build trust around it. We’re currently testing a video version that will be available on iPad, where you can discuss your nutrition with it, and the same avatar will show you how to cook anything from boiling an egg to making a crème brûlée.

Where does your conviction about home cooking come from?

I used to live in Sweden, and my roots are in Russia. In both, we still really value home-cooked food. It’s the top luxury you can afford, because you’re preserving nutrients, controlling the salt, sugar and fats you add, and it has not been reheated multiple times. It’s still the best way to feed your family.

Teaching kids and teenagers to cook also takes load off mothers. I learned young because my parents weren’t home and I had to feed myself, not because anyone enjoyed cooking with me. Mine are 13 and 15 now, and they cook for themselves and love it. Even travelling, they’d rather find a grocery store. They’ve got picky about fast food and delivery, and I think that fussiness is a good thing. We’re not against the occasional Five Guys.

Building iCook

Where do you think iCook stands apart from the rest of the nutrition space?

The ethical line for me is supplements. No way. The meta-analyses are clear. They don’t improve quality of life or long-term health, and they don’t extend lifespan. They’re a very small fix.

Meanwhile the food market isn’t developing. Waste keeps growing, and so does the energy and CO₂ cost of producing all of it, and people have no tools to do anything about that. The kitchen hasn’t caught up either. The last real upgrade to the home kitchen was the microwave. That was 1967. We've had sixty years of nothing digital between the fridge and the plate. There’s nothing digital that helps someone planning around a child’s allergies, a partner’s allergies, or their own diabetes.

People go to the supermarket with none of that support, and get bombarded by whatever influencers are pushing, three-for-two deals on sweets dressed up as recommendations.

Is that an outlier, or is it how the industry works?

The UK produces one of the best nutrition reports in the world. The Broken Plate clearly shows how much money we spend promoting confectionery, sugar, salt and sweet snacks — over 70% of all food advertising spend — while fruit and vegetables get less than 2%.

No matter what health campaigns we launch, there’s no intermediary between retailers and what they’re promoting to consumers. I’ve raised this with retailers at several conferences, and they tend to say it’s hands off — we’re selling what our consumers want — even though there’s a good deal of research on how marketing shapes behaviour. If the recommended portion size of a snack is 30 grams, but what’s sold is much larger packs, people under financial pressure may go for the bigger packs, which are better value and eat past the recommended portion regardless. As a registered nutritionist and public health expert, I find that hard to follow.

And that’s why I won’t pivot into supplements, because every other platform that took a scientific approach to nutrition has. All of them. My favourite influencers from five years ago are all selling them now, because there’s no money in recipes.

I want to build an ecosystem for people who still cook from real recipes, preserving the authentic ones and making them healthy with a balanced menu.

I also want guidelines for people who outsource the cooking, a nanny picking the kids up from school, so they know what to cook, that it’s balanced, and that the groceries are there.

What makes a founder credible to investors and retailers now?

I worked for a kefir company, Beautiful Dairy then and Beautiful Kefir now, on the R&D side, on bacterial cultures and their health benefits. I assembled a scientific advisory board, chasing down microbiome professors and top nutritionists, because we wanted a more credible slide for pitches to Sainsbury’s and Tesco.

The response was so low. People said, this isn’t for me. Now it’s the opposite. They turn up at pitch events asking to join your advisory board. The funding narrative has shifted to commercial traction plus institutional backing.

Redesigning the room

You’ve said startup culture — the pub, the late nights — isn’t built for founders with families. What have you seen work better?

My husband works for a large US tech company, and as a manager he isn’t allowed to do gatherings outside working hours. After 5pm or 6pm, everyone goes home.

I think that’s brilliant, and fair. When my kids were younger I was sending them to after-school clubs they hated, then calling my husband to do the pickup, because everyone was going to the pub and I had to go too.

I went to an event yesterday that ran from 7.30pm to 10pm. Most people there were men, or women without children. They’re 100% dedicated, but they don’t have small kids at home, and I barely see female founders who do.

It’s not that they’re less creative or won’t work hard. They have other duties that come first. I feel lucky to have teenagers now, otherwise I’d feel the FOMO constantly.

Last week I spent four days in Madrid at South Summit, where I pitched. The invite came a week before, so I shifted all my plans and paid for the travel and hotel myself. Nobody reimburses founders for that. I paid £500 for the flight and hotel to stand on a stage for eight minutes. The ROI calculus on pitch trips is brutal, and it filters out anyone who can't absorb that cost. And you’re doing that against 30-year-old guys with no families who won’t be expected to have kids for another fifteen years.

How would you redesign these events?

We don’t need female-friendliness. We’re fine with just having the same opportunities.

I’ve seen lactation rooms at major tech summits, which is great, but with the noise and heat I wouldn’t use one twice. A founder has meetings back to back, a pitch to prepare, and still needs to eat. Even on my own I was overwhelmed by the end of the day. Children’s rooms would be amazing. So would a free ticket for a spouse, or a nanny. I love how IKEA takes care of kids so parents can shop normally.

Beyond the events, childcare is the game-changer. In Sweden, parental leave is genuinely shared, and from year two the childcare is good enough that both parents can go back to work. The government pays the same for every child, whatever you earn.

What does doing everything right look like now, and why isn’t it enough?

The calibre of startups now is completely different from two years ago. With AI, we’re all learning faster what we’re expected to show up with. People who weren’t confident with a pitch deck now are, including me.

I prepared my cap table in a day, and that used to take several consultations with advisers. AI as a co-worker is brilliant for market research. If I meet an investor from Brazil, a proposal for that market takes hours, not a month, something solid on what happens if they invest and we advertise there.

It takes a lot of mental resilience. I’ve watched founders get overwhelmed by the sheer number of pitches and emails.

Scientific-minded people don’t like to chase. We want to be chased. That ego doesn’t help, and it doesn’t help with being visible either. I fought it for years, and I’m putting out more public content now, which has brought the ego down, because it forces you to explain things so people understand them.

Advice for the next founder

What would you tell a founder starting where you started?

Hold your ground. Nobody knows better than you do. If you believe in your idea, and you have the competencies to back it up, and you see the potential, don’t let anyone ruin it.

I've been told more than once that recipes are dead and ChatGPT replaces them. Those people have never had to feed a family of four with two different allergies on a Tuesday night.

I smiled and nodded. I don't do that anymore. I should have said: ask your mum. Is she cooking, or ordering delivery and buying meal kits? She’ll tell you she wouldn’t overpay £100 a week when she can do better herself. I once asked my mum why she still peels potatoes with a knife, and she said it’s because her dad did it that way.

Those things are what make us human, and delegating cooking entirely isn’t going to end well. Every founder needs to hold their idea firmly and be vocal about it, more vocal and bolder than I am.

I've learned to pick my moments. I have strong opinions on food advertising, supplement culture, and how the industry treats evidence. I'm more vocal now than I was two years ago, and I'll keep going. I work hard and overdeliver, but I have my own opinions and I’m not quite brave enough to put them out there. Maybe after a Series A. Just not yet.

Why does a Series A make it safe to be bolder?

Funded founders get read as visionary. Unfunded founders with the same opinions get read as difficult. I know that dynamic, and I'm navigating it, but I'm not waiting for permission to build what I believe in.

Is there anything you’ve read or listened to recently that you’d recommend to our readers?

I read Isaacson's Musk biography. Not for the success — for the failures. The scale of what went wrong and how much one person can absorb and still make decisions. That's what I took from it. What fascinated me was the scale of his actual failures, not just potential ones, and how much a person can process. He’s still across the engineering himself, knows the weight of every detail, and makes huge decisions anyway. We spend too much time worrying what people think of us, and as women especially, we worry so much we forget to do the job. I’d rather not spend that capacity on overthinking. So when people ask if I’m a feminist, yes, but I’d rather prove it through actions than volume.

I also love Michelle Obama’s Becoming, two ambitious people with similar but different ideas about what they want from life. That’s very much my husband and me, and our friends call us a power couple. I have strong family values, I love raising children, and I’d love more. Maybe not as many as Elon Musk.

This series is run in partnership with the Jessica Vollman Foundation, a non-profit founded to honour the legacy of the late CEO, founder and advocate for women in entrepreneurship: Jessica Vollman.

Halfway through our interview, Amber Vodegel’s 16-year-old daughter texted to ask for a lift. She was sitting in the same room. Amber was mid-sentence and didn’t reply, so a handwritten note arrived instead, with tick boxes for yes and no. Amber ticked yes, then broke off to explain it would have to be after five, because she had one more meeting.

For many female founders, this is what building a business looks like.

Amber has been building healthtech companies for the better part of two decades. She bootstrapped her first company, Pregnancy+, without investors or a network, while raising a family, before selling it to Philips in 2017. Pregnancy+ has since gone on to reach 150 million users worldwide.

In 2023 she started again with the women’s health platform 28X, backed by £1.2 million from the Philips Foundation and angel investors — and by a list of the things she wishes she’d known the first time.

Capital, networks and time constrain every high-growth female founder in the UK. Amber has worked around all three twice.

What we discussed

• How she bootstrapped Pregnancy+ without knowing that grants or investment existed, sold the business to Philips and saw the product go on to reach 150 million users worldwide.

• Why she started again after the Philips exit, and the book that shaped what she built next.

• How 28X stays free for every user.

• Her proposal for a basic certificate before you can register a company.

Lessons for policymakers

• Reform grant assessment panels to include entrepreneurs. Amber argues that assessors, often drawn from research backgrounds, are trained to reward long feasibility studies over founders who say a problem can be solved quickly. A panel split between an entrepreneur, a researcher and a policy specialist would change who gets funded, and for what.

• Exempt care costs from benefit-in-kind tax for scaling companies. Employer-paid childcare, school care and elderly care are taxed as a standard benefit, so the employee pays income tax on the value and the company pays Class 1A National Insurance on top. She would exempt high-growth companies from that treatment on care costs.

• Give Companies House a teaching job. There’s no basic knowledge requirement to register a company. Plain-language videos on what alternative capital is, which sort suits which company and what grants exist would point founders in the right direction.

Lessons for founders

• Constraints buy you the right to say no. Bootstrapping let her turn down beauty-product sponsorships aimed at pregnant users, which she doubts she could have done under investor pressure.

• Ask what your kind of capital demands in return. Venture and private equity suit some businesses but not others, and most founders treat money simply as money.

• Architecture is a business model decision. Because 28X runs on-device with no cloud back end, there’s no incremental cost per user. One woman or 100 million costs the same, which is what makes free viable.

The founding story

You bootstrapped Pregnancy+ without investors or a network, sold it to Philips in 2017, and the product has since reached 150 million users worldwide. Six years later, you started again from scratch. Why?

I bootstrapped the first company simply because I didn’t know funding existed. I made some money, put it back in and made some more. I never joined a network or went to an event. I worked in complete isolation. I had no idea what I was doing in terms of preparing for an exit, structuring the company or getting a grant. I just wanted to make the best product there was for our users.

I learned a lot at Philips over the years that followed the exit. We integrated with Medicaid in the US and went deeper into low-income countries, where Philips is excellent.

After I left Philips, I read Rutger Bregman’s Moral Ambition. It argues that the biggest waste of our generation is ‘a waste of talent’, that our best mathematicians are spending their brains on dopamine triggers and TikTok feeds when they should be working on climate change and cancer. Everyone follows the money. Everyone wants to be the next big tech entrepreneur, no matter the effect their product has on society.

Today’s tech products are often venture-backed, which means the user pays one way or another. The best period trackers cost £4 to £8 a month. That’s fine for a small percentage of affluent women. Unfortunately for a lot of women and girls, it simply isn’t affordable.

Imagine a 14-year-old girl from a low-income household where English isn’t her first language, who wants to track her period and understand her own body. She downloads an app and immediately hits a paywall. We treat that as normal. And if she doesn’t pay, her data gets shared instead, or she’ll be exposed to endless advertising. One way or the other, investors need to see a return on their investment, so the user becomes ‘the product’.

So I thought: let’s rethink the healthtech space. We’re building a women’s health app with no subscription and no data business behind it. We don’t use cloud servers at the back end, so we can’t access your data ourselves. We give women back the power over their own data, without having to pay for it. Your body. Your data. Your choice.

If there’s no subscription and no data being sold, what pays for 28X?

The good news is that if you don’t have cloud servers, you don’t have costs that scale. Our core technology costs don’t rise in proportion to the number of users, which fundamentally changes the economics. You download the same binary from the store onto your phone, and everything happens on-device. You can still share your data with your GP if you want, or back it up to your own Google Drive or email. It’s your own safe, not mine. There isn’t a meaningful incremental infrastructure cost attached to each additional user.

That’s the first part: we’ve dropped the cloud, and the costs that come with it. The second part is sponsors, the same way I did it with Pregnancy+. You can make tens of millions a year from ESG sponsorships that actually mean something. Educational partnerships, for example, and health solutions and products that are genuinely useful to women.

What did bootstrapping force you to do differently?

I think you become more creative. There’s less pressure, and you end up in a more ethical position, too. At Pregnancy+, before we were acquired, we were offered certain sponsorships that I didn’t think were helpful. Beauty-enhancement products, mostly. When you’re pregnant, you should be focused on life — your mental health, your social network — not worrying about your appearance. We were offered substantial sponsorships like that quite early on. If we’d been VC-backed, I don’t think I could have said no.

Bootstrapping gives you much more space to make the right decisions. You also have time, so you cut fewer corners on security and privacy. If getting it right takes a month longer, there’s nobody telling you it has to ship by June.

Venture capital is wonderful for certain businesses, but definitely not for all of them, and the same goes for private equity. There’s very little education on that. People see money as money. They just think, “I need money”, without knowing what it means or what the consequences are.

What she didn’t know

You built Pregnancy+ without knowing that funding, grants or networks existed. What needs to change so the next generation doesn’t have to learn it the hard way?

I think entrepreneurship should be on the school curriculum. It’s going to be huge, because in this new world, everyone can build anything with AI. You’ll see lots of micro-companies starting up, each doing something very specific for one area.

My co-founder’s son is 13. Over a few weekends he built a game — AI lets you do that now. He can now invite his school friends to play, and in return they invite another school to play against them. So now you’ve got kids in Guildford schools playing a game together, and they like it more than playing something global like Roblox, where you have no idea who you’re playing against. You’ll get far more companies built for a local purpose, and some of them will turn out to be valuable.

But those kids, in those schools, need to know what’s actually available to them in the UK, and how to access it. That’s where it has to start.

So where should that information come from?

Look at what happens when you register a limited company with Companies House. There’s no basic knowledge test, the way there is for almost everything else you’re allowed to do. You need a theory exam for a driving licence. You need a PADI qualification before you’re allowed to scuba dive. But you can set up a company with no understanding of what you’re getting into at all.

Companies House could put out a series of plain-language videos. One in seven adults in England has literacy skills at or below the level expected of a nine- to 11-year-old, so explain it at that level. This is what alternative capital is. This is what’s good for which kind of company, and why it’s not suited to others. These are the grants you can apply for. Just enough to point people in the right direction.

I’d push for some kind of basic certificate. At the moment starting a company is like having a child. Everyone says good luck, and that’s the training.

What needs to change

You want childcare treated differently for growing companies. What’s the thinking?

If you’re a growing company with a good revenue track record, you’re bringing new money into the country, because you’re selling a product or a service. Those are the building blocks of society. You have surplus value.

I think companies above a revenue growth threshold — set it at 20%, or tier it — should be able to claim care costs back, whether that’s childcare or elderly care. I think that makes total sense. With all of these suggestions, there are always people who find ways around the system, so it’s difficult to do perfectly. But there must be something we can do, especially for women, around childcare that you can claim as a benefit.

You’ve also mentioned proposing a minimum proportion of government spend going to UK-based companies. What’s the example that makes the case for you?

Take AI scribe software for GPs, the kind that allows a doctor to focus on the patient rather than typing throughout the consultation, with the conversation captured and the relevant information added to the system automatically.

When public bodies procure technology like this, I think there should be much more consideration given to where the economic value ultimately goes. It is not just about where a company employs people, but where the ownership sits, where the money flows and whether we are helping to build capability in the UK.

Supporting British companies should be a factor in those decisions. There should be a requirement for a meaningful proportion of public and grant-funded spending to remain in the UK. If a company has received an Innovate UK grant, for example, there should also be an expectation that a reasonable share of that funding is spent with UK subcontractors and suppliers. I think Innovate UK is already relatively good at encouraging that.

Quotas draw pushback. How do you answer it?

You do get that, and I see it. But maybe you frame it as either UK-based or female, like setting a few really important pillars, rather than one narrow box. That’s the only way to get people moving in that direction, if there are mandates around where the money goes.

I don’t think you can just hand money out. But you can set guidelines where a company has to tick some of those boxes — UK-based, female-founded, whatever the pillars are — so it falls into one of those categories rather than being excluded outright. If you do it that way, you don’t end up with a group of people against you.

You’ve said government grant programmes, including Innovate UK, tend to favour applicants from research or institutional backgrounds. What did you mean by that?

I didn’t know any of this until I won a grant myself. I’m getting better at them now, but the mechanism is revealing. In my experience, the process can feel heavily weighted towards research and institutional thinking, and I’d like to see more successful entrepreneurs represented on assessment panels.

So when I wrote my first submissions, I’d say something like: I only need this much money, and I can get you that much return, by just doing these three things. Very simple. They didn’t like that wording at all. They really don’t like it when you say it’s easy, that you can just do it, because that’s seen as far too entrepreneurial.

Every grant writer told me the same thing: say the idea is high-risk, say you have no idea whether it’ll work, ask for a feasibility study, then another, possibly a third. And always include well-known research partners as part of your funding proposal. That’s the setup and wording most assessors prefer.

What would a better process look like?

I’d want assessor panels to include a successful serial entrepreneur alongside someone from a research background and someone from policy. The three of them vote together. Right now, successful serial entrepreneurs are missing from these panels entirely.

Closing thoughts

What’s one thing you’ve read, listened to or come across recently that you’d recommend to readers?

I was really impressed by an article about Thuria Wenbar in The Times.

I know her personally, and I think the best businesses are built on strong ethics and a genuine desire to solve real problems.

She wanted to help improve access to healthcare, so she founded an online pharmacy and digital healthcare business. Through the service, people can complete a clinical assessment online and, where appropriate, receive treatment following review by a qualified healthcare professional. The information can also be shared with their GP to help support continuity of care.

I used the service this morning to request travel health support ahead of a safari in a high-risk malaria area. I completed the assessment online, and the relevant information was shared with my GP.

She’s a female founder who set out to improve the system and built a business to do exactly that. What makes her story even more remarkable is her journey. She was an asylum seeker, fleeing first to Egypt, then Germany, before arriving in the UK at the age of eight. She is one of many extraordinary women who have come to the UK from around the world and built businesses that make a real difference.

When you read a story like hers, an asylum seeker, a doctor, now running a technology business, you might think, how brilliant is that? But I suspect there is still a tendency in the UK to ask why someone like her did not simply stay in medicine.

The point is that she has not stepped away from healthcare; she has found a different way to contribute to it. She is solving a real problem, building something the NHS did not have, creating technology that can support patients and clinicians, and developing intellectual property that could ultimately be exported and licensed into other health systems.

We should be much more comfortable celebrating people who use their expertise to build companies, create jobs and solve public service problems at scale.

This series is run in partnership with the Jessica Vollman Foundation, a non-profit founded to honour the legacy of the late CEO, founder and advocate for women in entrepreneurship: Jessica Vollman.

Grace Almendras-Castillo built a healthtech company in Toronto, raised most of its money from US investors and sold it to an American buyer. She could have stayed in Canada, where she had 30 years of relationships. Instead, she moved to the UK, where she knew one person, and started again.

The UK’s 5.5 million small businesses drew her in, and so did its finance sector, which she rates alongside New York and Singapore for rigour. Gifftid, the AI platform she’s since built, exists to help overlooked businesses get in front of investors, banks and corporates who could fund them.

Getting there, however, took work. Grace spent her first months in London entering rooms where nobody knew her name, letting the relationships build before she asked for anything.

It’s how she built her first company too — flying to Silicon Valley, New York, Boston and Chicago several times a month, convinced a founder had to be there in person to be taken seriously.

We spoke about what it takes to build in a market you weren’t born into.

What we discussed

• Building Self Care Catalysts and selling it, and raising most of its money in the US.

• Why she chose the UK to start again in her 50s.

• Why she thinks founders have to be physically present in the market they want to grow in.

• Why she thinks the funding gap for women is a problem of evidence more than bias.

• Her advice to founders raising capital, and how to tell when you’re pitching the wrong kind of investor.

Lessons for policymakers

• Judge small businesses on more than their accounts. A company’s accounts show whether it can pay its debts, and little else. So the things a founder does to keep the business alive — mortgaging the house, going years without a proper salary — show up only as evidence that they are a bad bet. Grace argues a bank could learn more from how a company operates than from its balance sheet. Financial institutions are not expected to change their current risk assessment models, but to curate more intelligence not typically included in traditional underwriting.

• Open up data in the sectors that don’t have it. Financial platforms building a finance product in the UK can plug into open banking and other public intelligence. Grace had no equivalent in healthcare, where the data belonged to patients, every use required consent and nobody in the industry would share.

Lessons for founders

• Be there yourself early on. You can hire a team who knows the market, Grace says, but at the start the founder has to turn up in person.

• VCs aren’t the only route to capital. Angels, super angels, revenue-based financing, debt and grants are all legitimate paths, but people don’t always think — or know — about them early on. Choosing the wrong type of investor for your stage, Grace argues, is a common misalignment.

• Match the investor to what you’re building. Investors have different theses, and Grace argues founders often read a bad match as a rejection. If you’re not offering deep tech or AI, Silicon Valley VCs aren’t interested.

• The people who back you early are rarely convinced by your pitch deck. They invest because they trust you — your integrity, your track record of delivering despite constraints — and that kind of trust, Grace explains, is hard to convey on a slide.

The founding story

You built Self Care Catalysts in Toronto and sold it to Alira Health in 2022. How did that lead to Gifftid?

The thesis of Gifftid came from decades of work. I started in the corporate world, at some of the largest pharmaceutical companies, then became an entrepreneur with Self Care Catalysts. That gave me several vantage points. Inside a large company, I’d worked with small suppliers, decided where to invest our capital and set up joint ventures.

Afterwards I wanted to work on behalf of small businesses. Large corporations have capital, infrastructure and people. They also depend on small suppliers who have none of that, and who struggle to win a contract long enough to plan or borrow against.

What was the shift from corporate life to running your own company like?

Self Care Catalysts was a turning point. Coming from a cushy job, I had to learn how to run a small business — put my own capital in first, then raise more, then build teams.

I started with my own, my family’s small capital, then friends, then people who’d never met me and backed me anyway because they could see I was trying to solve something in healthcare. The hardest part was doing it on limited funds and hiring people yourself, with nobody to help you screen them.

You do everything when you’re a start-up — building the team, building the product, selling, commercialising — and you’re always raising.

What got me through was my team. It matters enormously to have people behind you who back what you do with capital or commercial contracts, and they do it because of your integrity and your ability to deliver despite the constraints. None of that shows up in a pitch deck.

From North America to the UK

How did being on the other side of that — a small business being judged by investors — shape Gifftid?

I’m Canadian and I lived in Toronto, but I raised most of my money in the US. I had some Canadian investors too, but I commercialised the company in the US and it was acquired there. It was the natural market for it.

My commitment to small businesses isn’t about a big market on paper. It comes from having been one, and from working with so many that deserve more capital and capacity — because they can use it.

Founders are the first to take risks. So when a bank turns you down and calls you high-risk — well, yes, but taking risks is the whole job. It runs from mortgaging your assets to going years without the salary you deserve.

Those risks don’t show up on a balance sheet — financial statements only read your ability to pay. So much about a small business is illegible to funders.

So I built Gifftid out of that conviction. First I found people who believed in it — some were classmates from Oxford’s Saïd Business School. They didn’t share my experience, but they had conviction and said, “let’s work with you on this.” Others were investors from my last company, or friends who backed me without quite knowing what I was building.

One day a major investor texted me: “Grace, can we invest in your company?” She was asking permission, without knowing what it was. She’s my biggest angel investor today.

None of that shows up in a pitch deck — and a deck is easy to assemble now, especially with AI. An investor or bank could learn far more about a company from other signals. That’s what we do. We gather public intelligence and operational signals that never reach a balance sheet, and turn them into real-time evidence for small businesses.

Why did you choose to build Gifftid in the UK specifically?

Moving here was a risk. I’m in my 50s, and I left Toronto — home for 30 years, where I had all my relationships — for a country where I knew just one person well: my former chief of staff.

I moved because the UK is willing to change how it deploys capital. First, it’s one of the most rigorous and respected financial centres in the world, alongside New York and Singapore. Second, it has 5.5 million small businesses — where else? — and a whole ecosystem around them.

One lesson from before is that you have to be present where you want to grow — build relationships, get out and learn. So I listened. I walked into rooms where I knew nobody, and I was lucky to always be invited. Sometimes I’d wonder why I was there; usually it was simply to listen and learn. That’s what the UK has given me.

Your first company took years to find its market. Is building faster now?

I tend to enter markets that don’t exist yet and help create them. Digital health wasn’t established when I started. I was one of its early builders. More than 200,000 patients were using our technology, until we pivoted to serving the pharmaceutical companies who became our customers.

It was slow partly because the market didn’t exist, and partly because I was challenging the system. In healthcare, patients used to be recipients of care, left out of treatment decisions. The power sat with doctors, hospitals and big companies. Only now is that shifting.

That’s why the company was called Self Care Catalysts. The whole idea was to give patients their own data and let them make decisions with it. It took a long time.

Healthcare is also harder because the data belongs to the patient. Using it means getting consent, and back then the ecosystem wouldn’t share. Compare that with open banking here, which is well established across the UK and EU — a completely different environment for new ideas.

Raising money in the US, though, is fast. I closed my first external round after speaking on an innovation stage in Washington, DC, about the technology we’d built. The technology took enormous time and money then. I could build it in a much shorter time now.

Here the environment is ready for new thinking, and the infrastructure for generating data is far better. Back then I was building my own very basic natural language processing. What used to be slow for lack of data is fast now — and it helps that I know what I’m doing this time.

Being present where you want to build and grow

What would you say to a founder expanding into the US who can’t be there?

You have to hire a US team — people who know the culture, already have the relationships and understand the business you’re entering. But if you’re really early, the founder has to be there.

Building my first venture, I was on a plane to Silicon Valley, New York, Boston or Chicago several times a month. It was easy, though, because I was based in Toronto.

One thing about the US is that once people believe in you, they say “I like you, I’ll introduce you” — and they do it straight away. You get picked up and brought around fast.

In the UK, I identified advisers who might be aligned with what we’re doing and reached out cold. They came on board, hugely committed, and have been introducing us across the ecosystem. That’s why I believe in finding people who believe in you. You don’t have to ask them to work hard, they just do.

Are there programmes on the US side worth UK founders knowing about?

Springboard Enterprises is an all-women accelerator that started before accelerators were even in vogue. They do everything to introduce you — booking meetings, putting you on stage, working for the founders. They brought us to Washington, DC, the White House and Silicon Valley. That was years ago, so I can’t speak for how it works now, and there are far more programmes today than there were a decade ago.

Here, I’ve deliberately not joined an accelerator programme because I already had so much experience, and I’d rather spend my time building and finding my own supporters. Whether that’s the right thing or not, I don’t know. But I’m happy with the progress we’re making.

Gifftid’s thesis

Who are the invisible entrepreneurs, in practice?

We’ve identified and named thousands of what we call the modern SME — smaller companies solving problems for society, whether in climate, energy, healthcare or food and agriculture.

Those are businesses traditionally seen as fringe or niche, but when we look at their forward-looking signals, they’re the ones with high commercial potential. The limitation is that they’re not given capital.

Immigrant entrepreneurs are the hardest to fund. With no starting point, you’re constrained on every side — capital for the company, building a new life, even the cost of attending events. But they’ve already moved to another country. They’ve been through enough hardship that the difficulty of building a business is easier to absorb.

Does the same logic apply to women?

The logic applies, but I’m probably one of the few who’ll say it differently. It’s not that investors don’t want to back women. It’s that there isn’t yet a long track record of returns to point to. Women only started founding companies in real numbers a few decades ago, so the track record investors look backwards for barely exists. That’s the gap.

Women tend to be very capital-efficient — we can do a lot with little — but those signals aren’t captured anywhere. There’s nowhere for them to be recorded. That’s what we’re building — the infrastructure to make those signals visible.

Women and foreign-born founders building companies that tackle climate, energy, healthcare or food — those are the modern SMEs. We’re the first to name them as such.

Advice and closing thoughts

What’s your advice for female founders on proving themselves to investors when the track record and evidence they’re being asked for might not exist yet?

It’s always a combination. First is economics, because when you’re talking to an investor, they need to make their money back with good returns. Investors have different lenses and different investment theses.

If you’re pitching to Silicon Valley VCs and you’re not offering deep tech, AI or something that returns at least 50x or 100x, don’t bother. Find the investor whose thesis aligns with who you are and what you’re doing.

Second, VCs aren’t the only route. I didn’t raise from them — I raised from angels and super angels, then borrowed against revenue once I had it. When I hear, “why does only 2 per cent of VC funding go to female founders?”, I’d say a lot of it is misalignment. You’re chasing a type of capital that doesn’t see you the way it needs to make its returns.

Third, be yourself. I never changed who I am, but I learned and got better. Still, respect that you’re there to ask for money and have to prove you’ll return it on good terms. If VC money isn’t the fit, look elsewhere — debt, grants, blended or revenue-based financing, or your own money first.

It’s not just about the money, either. You have to think about the nature of the capital you’re seeking and whether you can work with that funder at all. It matters most with PE and VC, because they have their own thesis, and you have to be ready to absorb that. Even after a first or second meeting, you won’t fully know who they are.

As a first-time founder I learned this the hard way. I thought doing good for society was enough, but you also have to show them when you’ll land your first million in revenue — that’s what earns a VC’s respect. If you haven’t got commercial revenue yet, don’t chase VC money. Spend the time building instead.

The number one thing: learn how to build and run your business, then pitch it.

What’s one thing you’ve read, listened to or come across recently that you’d recommend to readers?

This is going to be self-promotion, but I wrote a book called Boundless: Every Day. In Spite Of. It’s about how I made my decisions to get to where I am, and I’ve built an executive programme to go with it.

This series is run in partnership with the Jessica Vollman Foundation, a non-profit founded to honour the legacy of the late CEO, founder and advocate for women in entrepreneurship: Jessica Vollman.

🏦 Mann Virdee, Head of Science and Technology

Britain had a regional banking system and then, after a century of mergers and acquisitions, didn’t. Almost exactly 190 years ago, Midland Bank was founded in Birmingham. As it grew, it acquired other local banks, such as the Union Bank of Birmingham in 1883, and moved its head office to London in 1891. By the 1930s, Midland was, by some measures, the largest deposit bank in the world. It’s now part of HSBC. Many other regional banks followed the same path. The Manchester and Liverpool District Bank, for example, was acquired by National Provincial Bank, which later became NatWest.

While Germany still has its Sparkassen-Finanzgruppe (a network of local, publicly owned savings banks), we have branch networks whose credit decisions are often taken far away from the borrower.

That’s the broad context for two independent research papers on the geography of finance, which have just been published by the Financial Conduct Authority.

One of these papers found no evidence of a positive correlation between the overall size of a region’s financial sector and how much its smaller firms can borrow. That is, deeper capital markets don’t necessarily flow through to founders.

Instead, the authors argue, what matters is the composition rather than the scale. Their research found that lending to SMEs correlates consistently with regional growth. Broader financial expansion helps only up to a point, beyond which the relationship turns negative.

This means that geography, which digitalisation was supposed to have rendered unimportant, persists. It’s proximity that still shapes both small-business lending and whether larger firms make it onto the London Stock Exchange, because trust and tacit knowledge don’t travel well. Investors also price risk more steeply with distance from London than their counterparts do with distance from Paris or Frankfurt. Both papers find that the benefits of a financial centre don’t spill over much beyond its immediate hinterland.

These are associations rather than proven causes, and the FCA cautions that these findings should be treated as indicative and considered alongside other evidence. But at a time when Manchesterism is top of the agenda, it’s worth exploring whether the ability of a founder in Warrington or Wakefield to raise capital depends less on the depth of British capital markets than on where the decision to lend gets taken.

Share

🧪 Philip Salter, Founder

Every disease is a policy failure. That’s the thesis of Saloni Dattani’s essay for Works in Progress. Her case is that we treat disease as part of the natural order, something to be endured, when most of it is really a problem we could solve — if we funded, organised and incentivised the work properly.

As she argues: “Medical innovation is one of the most valuable things that happens in a modern economy. Economists estimate that gains in life expectancy in the United States from 1970 to 2000 were worth about $95 trillion to Americans, with roughly half of that coming from reductions in heart disease mortality.”

However, we are leaving some of the greatest value on earth uncreated, not for want of ideas, but for want of a buyer. To put it bluntly, the market won’t serve patients too few or too poor to make a market. The malaria vaccine is a case in point: the science was largely there in the 1990s, but it sat for decades while half a million children died each year.

One answer is to incentivise the private sector. An advance market commitment (AMC) does this by design: donors promise up front to buy a proven product at a set price, stimulating the demand and certainty a firm needs to invest, then letting private firms do the capital-intensive work of trials, regulation and manufacturing. The pneumococcal AMC launched in 2009 is estimated to have saved more than 700,000 young lives.

Another part of the problem is testing, and here Britain has a genuine asset. Saloni writes about how the RECOVERY trial ran through the NHS during the pandemic and tested a dozen treatments against a shared control group at a fraction of the usual cost. It found a cheap generic steroid that saved hundreds of thousands of lives in months. A single national health system can run a trial across the whole country at a speed almost no one else can match — if we treat it as an engine of innovation and not merely a provider of care.

Everything above — and a lot more — is in Saloni’s essay, which is worth reading in full. On the one hand, it’s a heartbreaking realisation that the costs of bad policy are a matter of life and death on an epic scale. On the other, it’s a call to arms that shows how far innovation has taken us, but, more importantly, how far it will take us — and how quickly — if we get the funding and incentives right.

Subscribe now

📈 Sophya Mashkoor, Researcher

Entrepreneurship is suddenly looking a lot more attractive to British adults. According to a survey by QuickBooks, 66% of UK adults are considering starting a business or side hustle, an increase from 52% the year before. More interestingly, the share who actually intend to start one in the next 12 months has doubled to 30%. Almost half of adults surveyed say they earned money from a side hustle in the past year, and 45% of those people hope to turn it into a full-time business — though slightly more, 49%, are content to keep it as side income.

Is this a sudden surge in entrepreneurial zeal? Not quite. Respondents put starting a small business or side hustle top for potential returns next year on 26%, ahead of cash and savings (21%) and the stock market (11%). Only 13% think it is the riskiest of the options they were shown, below both crypto (32%) and shares (20%). Higher return and lower risk than equities is less a claim about entrepreneurship than it is a verdict on the alternatives.

The survey shows that earning more money is the single thing Brits say would most improve their satisfaction with life, outranking better health, relationships or more free time. Of those considering starting a business or side hustle, just over two-thirds would begin it alongside existing work.

You’d expect a lack of ambition to be the biggest obstacle, but Britain may have more entrepreneurial ambition than it is given credit for. What stops people is a lack of money. Asked to name their single biggest obstacle, 30% said a lack of savings or startup capital, ahead of 26% who said fear of failure. Six in ten say they would be more likely to start a business if financial guidance were easier to access. Asked what would most help them feel ready to fund a business, 32% chose a small grant, against 12% for a clearer funding plan.

In The Entrepreneurs Network’s own recent survey of UK founders, Cool Britannia, a similar pattern emerged. According to respondents, the strongest signal that a country is serious about entrepreneurship was tax levels and breaks on 68%, followed by access to capital on 60%. As it turns out, aspiring founders and established ones are asking for the same thing — easier access to finance.

Become a Supporter or Adviser

In 2019, a landmark review commissioned by the Treasury found that if women started and scaled businesses at the same rate as men, it could add £250 billion to the UK economy. The review made headline news — it had a catchy number, political backing and a list of recommendations that were meant to change the picture.

Seven years on, that picture looks much the same. Female-led startups still receive a fraction of UK venture capital, and the warm introduction culture that locks founders out before they’ve had a chance to prove anything still shapes who gets funded.

What happens when you have the skills, the experience and the idea — but none of the network? Ahana Banerjee had all three, but she didn’t know the right people to help her get started.

Despite being an Imperial physics graduate with a strong record of software engineering internships, she submitted 200 funding applications and only received one reply that wasn’t a rejection. That reply was Y Combinator, which, for her, changed everything. Because in the fundraising world, credibility is — albeit unfairly — built on who has already bet on you. From there, she built Clear, a skin-health platform that helps people track their skin, understand what affects it and make more evidence-based decisions. It is now used by more than 70,000 people.

There is, however, a second number in Ahana’s story, and it comes later. At the end of her first proper funding round, she went back through her CRM, added a gender column to 300 investor meetings and did the maths.

What she found changed how she thinks about everything that came before.

What we discussed

• Growing up between the UK, India and Singapore, and how that shaped her sense of purpose.

• The confidence it took to get through hundreds of applications and rejections, and what Y Combinator gave her that the UK funding ecosystem couldn’t.

• Why she ignored the VC growth playbook and spent five years on product fundamentals instead.

• How her first funding round changed her perspective on what it means to be a female founder.

• Why the UK’s warm-introduction culture shuts out the founders most likely to build something worth backing.

Lessons for policymakers

• Consider tax incentives for backing female founders. Ahana frames this as one of the few levers that could be strong enough to force change. Without a clear financial reason for investors to back female founders, voluntary change has been too slow. The data shows female-founded businesses are not a poor investment, but the problem is access to the capital needed to prove it.

• Address the UK’s reliance on warm introductions. Early-stage funding here runs almost entirely on existing networks, which structurally excludes founders without wealthy families or established connections, regardless of talent. More early-stage programmes willing to bet on individuals and demonstrated traction, rather than existing relationships or five-year exit plans, would change who gets to build companies in the UK.

Lessons for founders

• Know where your odds are highest. Across 300 investor meetings, more than 90% of what Ahana raised came from female-led decisions. She didn’t know this until she went back through her CRM after the round closed, but knowing it earlier would have changed her entire approach to fundraising.

• Don’t let someone else’s playbook dictate how you build. Most VC advice assumes a category with a clear leader and a known distribution strategy. Ahana had neither. She credits Clear’s survival to ignoring the standard advice to spend aggressively on growth and focusing on product fundamentals instead.

• Persistence outweighs a lot. Ahana submitted 200 applications and got 199 rejections or no response at all. What kept her going wasn’t the certainty she’d succeed, but the possibility of a high-potential reward — so she refused to read too much into the silence. The one yes was the only one that mattered.

The founding story

You studied physics at Imperial, then built Clear, a skin-health platform that went through Y Combinator and now has more than 70,000 users. How do you trace a line from one to the other?

I was actually very much set on academia and pursuing a career in physics when I started university. Before that, I’d had some rather formative experiences as a teenager. I spent most of my childhood in the UK until the age of 14, and then my family and I moved to India.

I am of Indian origin, but I wasn’t very culturally Indian at that point, and India is a country of very stark extremes. My life there was experiencing the utmost privilege — I went to an international school, most of my teachers and classmates were from the US, and my peers’ parents were ambassadors, CEOs of very big companies. I had come from a regular state school in the UK, and yet I was living in a country where I was experiencing privilege as I had never seen before, and also seeing poverty and suffering as I had never seen before.

And there was that added layer of being of Indian origin, not very connected with my culture — just this feeling of, what did I do to deserve this life? Why are there these girls who look like me, my age, begging?

It had a huge impact on me. I became aware of all the privileges I’d had, and I’ve always felt that my greatest privilege has been my education. I wanted to use that education to do something broadly positive in this world. At the time, I thought that would be through scientific research and academia, which is why I spent all my high-school years just studying.

When I was 16, we moved from India to Singapore for those last two years of school. I got into Imperial and was delighted, because my life plan was on track. But it was pretty much as soon as I got there that I had a bit of an identity crisis. The more I learned about the realities of a career in academia — the timelines involved before you see the impact of your work, the job insecurity — I realised this isn’t what I want. But I still felt this huge responsibility: what do I do with this education and these skills?

So what did you do with that — the sense that academia wasn’t the right path, without yet knowing what was?

I just started applying for internships. I was lucky to be in London, where there are lots of different companies and sectors. The first was in software engineering at a big investment bank — I’d never had any interest in financial services, didn’t really know what it meant to be a banker, but I was curious. By the time I got to my third year, I had quite a few internships under my belt across software engineering and banking.

Then a guy messaged me on LinkedIn and said he wanted to build a company — matching students looking for graduate jobs with employers — and needed someone who could code, had events-planning experience and was connected with HR people at big firms. I ticked all three. At this point, I had no idea what entrepreneurship was, never even remotely considered it as a career, but it sounded like fun, so I said yes.

That was my first foray into entrepreneurship, and a huge turning point. I realised that as someone with both a technical skill set and a generalist skill set, this was how I could maximise my impact. But it was a family-funded business — funded by the very rich father of the guy who’d reached out. Because of that, it set up some slightly weird team dynamics, and I knew I didn’t want to start my career with this specific company.

The question then was: how does someone like me access funding? I had the confidence that I had the skills to build a business. The thing I didn’t have was capital, and I didn’t come from a family who could support it. So my actual plan was: take the graduate job that pays the most, save up and use that to fund my own business one day.

I ended up with two finance job offers going into my fourth year — one in private equity, one in investment banking. Then Covid hit. My extracurriculars stopped, the part-time job fell away and I suddenly had way more free time. I thought: if I one day want to start a company, why not start now?

Where did you start?

I started with the thing I knew nothing about, and that was the investor side. I submitted about 200 applications that summer with a pitch deck and got 199 rejections or ghosts. The one that didn’t was YC. They became investors in Clear when it was still at the idea stage, while I was still studying full-time. The condition was that I commit full-time, which meant dropping out of my master’s programme, and I made that decision at the start of 2021.

Off the back of YC, we did our first pre-seed round, which was about $900,000. Because that first fundraise was such a traumatic experience, I made sure that if I ever raised again, it would come from a position of strength, not need. I kept the team very lean — just me, then a CTO a year later. I still write code five years in, and it’s just the two of us full-time.

We’ve spent the first five years really focused on product fundamentals, tech and retention, de-risked a lot of that, proven organic growth and turned on monetisation at the end of last year on both the consumer and B2B side.

The turning point

What kept you going through 200 funding applications?

This is either a personality trait or a flaw, depending on how you look at it, but I’m a huge believer in focusing on what’s within your control and not caring too much about the outcome.

When I was applying for internships, I did about 100 applications per week. I was already so accustomed to dealing with rejection that it just didn’t bother me. I’m very good at delayed gratification. I do things where, in the moment, I’m wondering if it’s doing anything, but until I get the one result I need, I just keep going.

And at the time, I still knew so little about the world of investment that I felt: I’ve already made the pitch deck, it’s just a matter of answering the same questions in a new form. It’s relatively little effort per additional application, with a potentially large reward.

What did getting into Y Combinator change for you?

I’m going to sound overly dramatic, but it was life-changing, and not because my skills improved overnight. The thing that changed was the credibility.

I’d been to a top university, done a serious degree, interned at prestigious investment banks, and yet I couldn’t get a meeting with any UK investor. I wasn’t in the networks. I didn’t know where to go. Even with VC investors, the whole industry is so reliant on warm introductions. If you have a genuine cold start, it’s very, very difficult.

There’s also this constant advice to start with friends and family. But when you’re a university student, unless you come from a wealthy family, you’re not getting funding from friends and family.

When it comes to raising investment — even just getting the meetings — I can say with a high level of confidence that the only reason I’ve raised any capital since is because YC was the first investor.

YC has a reputation for pushing hard on speed and growth. Did that ever create tension with how you wanted to build Clear?

There is something to be said for that. US investors do expect more speed, and because most of my investors fall into that category, with a more traditional VC mindset, that doesn’t completely align with what I think is best for my company.

I have always believed that with what we’re building, there is no market leader in our vertical. I want to build a billion-dollar business in this category, but it has not been done before. So, in terms of what that looks like from an actual product perspective and what strategies work for distribution, we don’t know. There’s not an exact playbook we can follow.

Therefore, in the early days, had I just haemorrhaged that initial funding round into marketing, which is what a lot of VC investors would expect, I can tell you very clearly there would be no Clear today. The business would have died a long time ago. So I purposefully didn’t do that.

Maybe the consequence I’ve paid is the narrative of — it’s been five years, why has the business not grown to a gazillion users? But the answer is that I wanted to invest in product development and tech to de-risk the fundamentals, to build a business that is profitable if it needs to be. And now we can take money and put it into growth, and this is the time to do it.

Against the grain

You raised the gap between university and the funding ecosystem — the idea that early-stage capital is so reliant on angel networks that young founders, almost by definition, can’t access it. What would help fix this?

This is a very utopian goal, but investors should not be so reliant on warm intros. I get it, I understand that they receive more than they can possibly deal with. I also receive more emails than I can possibly deal with, but I deal with them.

Only leaning into your immediate community is what causes the bias in the first instance and is what objectively makes it harder for founders like me to get in. And you see it in the numbers because there are also fewer investors who look like me.

I’m really against this warm introduction culture because it’s what makes it so easy for some to fundraise and so difficult for others, and there’s very little correlation with how good a builder you are.

Were there barriers you faced as a young female founder that you didn’t expect going in?

My opinion on what it meant to be a female founder drastically shifted after that first funding round.

Going in, I was absolutely the kind of person who took the view that it doesn’t matter what your race, gender or age is, it’s about what you can do. Up until that stage, I actually thought being a young brown woman was a net advantage. I can chat about skincare far easier than your average 40-year-old white male software engineer, but I’m as good at writing code, so net, I’m the right founder to build this business.

I have a tangible advantage. I’ve always felt there was a huge amount of founder-market fit, and I think investors who do invest recognise that I have the correct skill set, background and understanding of the industry to build the company that I’m building.

The first funding round totally shifted my perspective. I had my standard CRM to manage all the meetings I was having with investors. That first funding round — I’ve glossed over this — was about 300 meetings in a very short time frame. It was awful.

When I say it was traumatic, it was just meeting the worst people, saying all kinds of things I had never had said to me before, which is where the gender element comes into it. I had, up until that point in my life, prided myself on my academic capabilities. That was a large part of my identity. And then there was this notion that I was stupid. There were investors who did not believe that I was the one who had written the code for what I was building.

There are also some specific examples that have stayed with me. I knew a guy who was a few years younger than me who had dropped out of high school to build his company. He looked like he was high all the time, spoke really slowly and didn’t ever answer a question directly. But investors loved the persona. This worked so well for him as a genius dropout. If we’re being real, he did not prove anything. Nothing about him said genius except for his cool dude persona. I’m not a cool dude. I get it. But if we look at actual credibility, I had a lot more than he did coming into this. Yet the whole narrative around the fact that I’d dropped out was perceived very, very differently.

I’d been told I don’t look like someone with a physics degree. I was genuinely having to send screenshots of my code commits, basically do live coding interviews with investors, to prove that I was the one who had in fact built my own company, despite having several software engineering internships under my belt, despite having a physics degree from Imperial.

The thing that validated the bias point was the CTO I hired a year after starting the business. He also has an Imperial physics degree. We have the same qualification. He is a white man. No one has ever asked him if he can code. No one has ever challenged the notion that he is a software engineer. I have had it challenged multiple times.

The case for change

How did that experience change how you approached fundraising?

I got to the end of that round and thought — is it just me, or are a lot of my investors women? I went through my CRM and added a gender column. Of the roughly 300 meetings I did, about 75% were with male investors and 25% with female investors, which is roughly expected.

But of my actual investors, over 80% were women. And in terms of dollars invested, accounting for the size of each investment, over 90% of the dollars I’ve raised were led by female investment decisions. From a sample size of 300, that’s a striking pattern in my own fundraising data, and it isn’t trivial. Had I known that before I started fundraising, I would have strategised very, very differently. I am strategising differently this time.

Female investors rarely asked me to justify the size of the skincare market. With male investors, it came up repeatedly — often in ways that revealed how unfamiliar they were with the category.

Fundamentally, after that first funding round, my opinion totally changed. I discovered that it is, in fact, considerably harder to raise funding as a woman, and it doesn’t matter how talented you are.

This was before I’d seen any of the 2% statistics, before I’d read the Harvard Business School study about promotion versus prevention questions in investor pitches. I knew none of that going in. It was only after I looked at my own numbers that I came to the conclusion that there are indeed systemic barriers for women trying to raise capital. I actually think it does women a disservice not to acknowledge them.

There are founders who continue to say it doesn’t matter — if I could do it, anyone could do it. That’s just not true.

If you could change one thing about the environment for high-potential female founders in the UK — on the policy side — what would have the biggest impact?

Explore whether targeted tax incentives could broaden who receives early-stage capital. I’m offering this as a provocative policy avenue rather than a fully formed proposal, but so far voluntary commitments alone have produced limited change.

Female founders have such difficulty accessing capital in the first place. If we just had more capital to prove what we’re capable of, demonstrate those returns and put more case studies in front of people, that could start to challenge these systemic biases.

I don’t think any of the people I spoke to who asked if I really had a physics degree, or if I could really code, are genuinely racist, sexist or ageist people. They’re just pattern matching. But when people said this to me, I said to their face: so tell me, what does a physicist look like? It just makes them uncomfortable. It makes them think.

And it makes me feel like if there were more examples of women with technical backgrounds, female founders building businesses — just for people to subconsciously re-evaluate what they think a founder looks like — that could help.

What’s one thing you’ve read, listened to or come across recently that you’d recommend to readers?

Black Box Thinking by Matthew Syed. The core idea that’s stuck with me: successful people don’t fail less — they just learn from failure faster. As a founder, that’s reframed how I think about growth.

The goal isn’t to have the perfect plan; it’s to run the right experiments, read the signal quickly and be honest enough to kill what isn’t working. I wrote a bit more about how I’m applying this at Clear here.

This series is run in partnership with the Jessica Vollman Foundation, a non-profit founded to honour the legacy of the late CEO, founder and advocate for women in entrepreneurship: Jessica Vollman.

🥼 Mann Virdee, Head of Science and Technology

Graphene was first isolated in 2004 in a lab at the University of Manchester, work that won Andre Geim and Konstantin Novoselov the Nobel Prize in Physics. But within a decade, more than 7,500 graphene patents had been filed worldwide — only about 50 of them from the UK. By contrast, more than 2,200 were from China and over 1,100 were from South Korea, with Samsung alone holding hundreds. The Chancellor at the time, George Osborne, insisted that technology invented in the UK should also be developed here. It’s a problem that keeps repeating.

It’s that kind of failure that sits at the heart of a new report, Science: A New Golden Age, published last week by President Trump’s science and technology adviser Michael Kratsios. It’s the boldest attempt to reimagine the US research ecosystem since Vannevar Bush’s famous 1945 report Science: The Endless Frontier.

One of the core arguments in the report is that discovery without domestic manufacturing can leave countries paying for the research while others capture the returns. Or, as I wrote for this series last year:

“process knowledge is being lost by offshoring supply chains, which in turn harms countries’ entrepreneurial ecosystems. The process of building, iterating, innovating, and improving manufacturing gets lost, and it’s just as important as the ‘Eureka’ moment in the lab.”

Britain could view the report as a mirror of what has happened in our own science and tech ecosystem.

Another key part of the report is its selection of the funding-mechanism experiments that Britain’s own metascience community has championed for years. That includes golden tickets (letting one reviewer back one idea that a review panel has rejected), fast grants decided in 48 hours, and portfolio-based allocation run like a venture fund. The report calls for empowered metascience units with the authority to run randomised trials on grantmaking, and – interestingly – holds up the UK’s own Metascience Unit, established in 2024, as a model to emulate.

Alongside this, there are proposals for increased direct-to-individual funding that sidelines universities, pressure to reduce research overheads, an AI-first mandate, and a proposed rule giving political appointees more sway over grants. But it’s a tough environment for such proposals; A New Golden Age arrives at a time when the trust between scientists and government has frayed, amid grant terminations, funding cuts, and reductions in agencies such as the National Science Foundation and the National Oceanic and Atmospheric Administration.

Share

🔧 Ian Ng, Researcher

“Can AI build a train?” asked Andy Burnham in his latest push for technical careers. Since ChatGPT took off in 2022, plenty of research has looked at AI’s impact on the labour market, and the common assumption has been that skilled trades are more immune to it than desk jobs. However, Google’s latest research complicates the picture.

Google’s first AI & Economy ATLAS report looks into how users interact with its AI tools and finds that AI use is not limited to desk-based jobs. It also assists workers in physical and manual occupations on adjacent tasks, acting as a “hands-on collaborator” for diagnostics, troubleshooting and real-time learning, with multimodal AI use over two times higher than the overall usage rate among automotive service technicians and mechanics.

The report compares AI usage shares to employment shares across major occupation groups. As you’d expect, AI adoption is the most over-represented (relative to employment share) in “computer and mathematical” and “business and financial operations”. While a correlation can be drawn between higher manual task shares and less AI usage, some high-manual occupations still stand out where AI usage was observed — in tasks complementary to manual work. For example, 44% of industrial machinery mechanics’ tasks are manual, but AI use is still observed in non-routine cognitive tasks such as analysing test results and machine error messages. That’s the key distinction — the routine, hands-on part of manual work is largely untouched by AI so far, but the non-routine analytical part of the same job isn’t.

It’s not the doom and gloom of AI replacing human beings either. Britain faces a well-documented labour shortage in key sectors crucial to the net-zero transition, in particular among automotive technicians and mechanics. Automotive technicians and mechanics are in mostly manual jobs, and the report observed AI being used for testing vehicle components and inspecting parts for wear. It happens to be one of the sectors where Britain faces difficulty in filling vacancies, as the lack of EV-qualified technicians is stalling Britain’s transition to EVs. The training pipeline is not keeping pace with the technology, and instead of replacing technicians, AI could actually assist technicians with diagnostic work. Another occupational group where high AI usage was observed is electrical and electronics repairers, recently included in the Migration Advisory Committee’s Temporary Shortage List.

Burnham is right that Britain undervalued technical routes. But the fix shouldn’t rest on the idea that trades are somehow immune to AI — they’re not, even if the disruption looks nothing like the white-collar version. AI was not built to plug Britain’s skills gap, but we could use some AI to plug that gap and help train the next generation of technicians.

Subscribe now

🎢 Philip Salter, Founder

There’s no getting away from it: being an entrepreneur is a risky business. Fewer than four in ten new UK businesses make it to their fifth birthday.

But we know that society benefits from this risk-taking. Entrepreneurship is one of the most powerful engines of productivity and progress. Throughout history, entrepreneurial ventures have transformed societies, lifting people from subsistence to prosperity.

Nordhaus estimated that innovators capture only about 2.2% of the social surplus their innovations create. The rest flows to everyone else. But we aren’t the best in the world at building a society around this fact. As Richard Browning, founder of Gravity Industries, argues:

“Culturally, we are far more risk-averse than places like California, often pricing ambition out of existence before it even starts. To compete globally, we must shift from an insular mindset to a more outward-looking, non-ministerial view of global opportunity.”

If Browning is right that we price ambition out too early, it’s worth asking how early. Writing today, Elle Smith argues that risk aversion starts in childhood — and that a freer upbringing might be what produces the risk-takers an economy needs. So what does the evidence suggest?

Ross Levine and Yona Rubinstein have found that it is the combination of high cognitive ability and a streak of teenage rule-breaking — skipping school, minor mischief, a taste for risk — that predicts who becomes an entrepreneur and who earns more for it. A Swedish cohort tracked for nearly four decades replicated the pattern, though only for modest rule-breaking rather than serious crime.

The strongest signal of all is having an entrepreneurial parent, which makes a child around 60% more likely to follow suit; and because the effect shows up even among adopted children raised apart from their biological parents, the evidence points to role-modelling, not genes alone.

What we can’t yet say is that giving children more freedom makes them entrepreneurs. The logic is inviting: free and risky play builds resilience, self-regulation and self-efficacy, while over-protective parenting seems to erode them. Self-efficacy is one of the best-evidenced ingredients of entrepreneurial intention we have. But no one has yet traced a free-range childhood through to who actually starts a business — the direct evidence isn’t there yet.

I don’t want to rest too heavily on anecdata, but in a recent chat with a large group of some of the most impressive people I know, nearly all of them said they’d had a great deal of freedom growing up. It’s a small sample and correlation isn’t causation and all that, but it deserves more thinking about.

Become a Supporter or Adviser

🔑 Philip Salter, Founder

When Stripe Economics reported a surge in American solopreneurs last month, it raised the question: is the same thing happening in the UK? According to the official numbers, no. The ONS says business creations fell 8% in the first quarter, and registrations have been flat for two years. Britain, apparently, isn’t starting up.

However, Stripe dug into the UK data and concluded the boom is almost certainly happening here too. Its UK sign-ups have doubled in two years, tracking the US and France almost exactly, and since 2024 they have pulled sharply away from Companies House incorporations.

We already know Britain is bad at counting this kind of activity. And what it does count doesn’t join up. Companies House doesn’t count sole traders, the ONS only counts businesses registered for VAT or PAYE, you don’t need to register with HMRC at all until you’re turning over £1,000 a year, and the Labour Force Survey — the survey meant to pick up everyone else — has lost roughly half its effective sample since the pandemic, was suspended in 2023 and returned as a shadow of its former self.

France and the US can see their versions of this shift because both keep dedicated statistics on solo and micro-businesses. France’s are the cleanest, because registration there is near-universal from the first euro of activity and the records join up.

We recently set out a fix for this in The Master Key, with Xero and Enterprise Nation. It proposes a Unique Business Identifier paired with reusable, verified digital credentials — one key a business carries across HMRC, Companies House, banks and local authorities. Its main purpose is to cut admin and fraud, but it would also let departments join up records they currently hold separately. It wouldn’t capture everyone, unless we required or incentivised the smallest side-traders to do what France does, but it would get us most of the way there.

Policy is built on data, and this gap calls into question a big part of the story we tell ourselves about declining business dynamism. The claim that entrepreneurship is in decline in Britain, creating fewer firms, with less churn and an ageing business base, rests almost entirely on data that counts employers and registered companies.

A country that can’t count its founders doesn’t know how to support them. Right now, Britain can’t count.

Share

📈 Ian Ng, Researcher

High taxes, red tape, gaps in late-stage funding, and a Westminster that doesn’t understand entrepreneurship have become common complaints from founders since we started surveying them a year ago. We published our fourth edition of the Entrepreneurs Survey this morning, and the results tell the same story again.

Most founders continue to find it difficult to scale a business in the UK. Our survey also found that nearly 80% of founders think the government does not understand their needs, and when asked which of the political parties they trust most to understand their needs, the most popular option goes to “none of the above”. Most say it is difficult to raise investment and hold a negative view of the level of taxation and regulation in the UK. You can see the full results here.

In spite of all the hurdles, entrepreneurship in the UK continues to be alive and kicking. Entrepreneurs remain optimistic about the year ahead for their business, and many would still encourage others to start a business of their own. More than 40% of founders expect to increase their headcount over the next 12 months.

As Keir Starmer bid farewell with his final PMQs today, his successor will inherit a mixed bag of policies impacting entrepreneurship: a higher rate for employer National Insurance, of course, but also AI growth zones, legislation for regulatory sandboxes and the Industrial Strategy.

What founders are asking for has stayed consistent, and so have the answers they gave in our surveys. They look for certainty, not months of flirting with key taxes ahead of the Autumn Budget, and for less bureaucracy. None of this requires reinventing the state, just a government willing to act on what founders have been telling us for a year. The next Prime Minister can be the one who finally listens, or the one who gives us the same results to report again.

Subscribe now

🔬 Dr Florence Young, Senior Public Opinion and Policy Officer, Campaign for Science and Engineering

The pioneering Research and Development (R&D) occurring across Scotland and Wales is expanding our understanding of the world, tackling society’s biggest challenges and bringing benefits to local communities. But this work, and the organisations doing it, are not well recognised.

Campaign for Science and Engineering (CaSE) works to champion R&D as a political and societal priority, including by exploring how the public think and feel about R&D, to help the sector make R&D matter to more people.

Our latest public opinion study – Public Attitudes to R&D in Scotland and Wales 2026 – confirms a consistent dilemma we see across the UK, that public support for R&D is broad but shallow. If our sector continues to feel distant and intangible to voters, it risks being deprioritised by politicians.

On the one hand, the vast majority of people value public investment into R&D and like the idea of R&D being present in their communities. Around three quarters think it is important for their devolved government to invest in R&D in their nation (77% in Scotland and 72% in Wales) and most people (66% in Scotland; 59% in Wales) would feel proud if their area were to become well known as an important hub for R&D.

But despite this, Scottish and Welsh R&D suffers from very low levels of awareness, and its benefits feel vague and hard to articulate for many.

Some 87% feel they don’t know much or anything about the R&D being done in Scotland, with an even higher proportion in Wales (91%). And people struggle to immediately identify the benefits of R&D, whether for themselves, their local area, or Scotland or Wales as a whole.

The sector cannot be complacent and must act now to deepen connections.

Our Advocacy Toolkit points to place, purpose and participation as powerful connection points. The R&D sector has a positive story to tell and CaSE is working closely with our members and the wider sector to make R&D more visible in communities, convey how it is contributing to the public’s priorities and highlight how the public can get involved.

Become a Supporter or Adviser

🪖 Mann Virdee, Head of Science and Technology

Navigating the world of defence procurement is notoriously difficult as a small firm. The new Defence Investment Plan, published yesterday, seeks to change that. I’m sure many entrepreneurs will welcome this, while others will naturally be more sceptical and will want to see it to believe it.

The headline figure is an increase in defence spending of £15 billion, funded from cuts to some road and energy projects – perhaps the last places such funding should come from.

The Plan explicitly frames defence as an engine for growth. Underneath some of the larger stories about submarines, sixth-generation fighters and drones, entrepreneurs will note that government is trying to become a better customer. There’s a commitment to spending an additional £2.5 billion through SMEs by 2028, representing a 50% increase in direct spend. That will be through the Defence Office for Small Business Growth, which was created at the start of this year.

Alongside this, there’s UK Defence Innovation, backed by £1.6 billion. That funding will be used to work with startups, scaleups and spin-outs to enable innovation in defence at ‘wartime pace’. The Plan also states that it aims to deliver ‘the next UK Defence unicorn’, with up to £100 million in accelerated contracts for British tech firms that have had limited or no current business with the Ministry of Defence (MoD).

This procurement-as-de-risking is one of the key ways the MoD can support entrepreneurs. When government sends clear demand signals and commits to buy, it helps SMEs cross the valley of death that kills so many promising firms between prototype and market. It’s the same logic that underpins Advance Market Commitments.

The last SME target, 25% of MoD spend, was technically met – but this was mostly through indirect spending via prime contractors. Direct spending with SMEs has remained around 4-5% (compared with 25% in the US), and as the government has noted, this figure has been falling. So this signal is welcome, but whether it will survive the transition to the next Government and actually deliver remains to be seen.

Subscribe now

📈 Philip Salter, Founder

For decades, the official figures have shown the British economy is barely moving — a percent or two of growth in a good year. Yet, the ingenious entrepreneurs I meet week in week out, building incredible things at an astonishing rate, are suggestive of a more optimistic reality.

In When GDP Misleads: Inferring Living Standards from the Value of a Statistical Life, Stanford economists Philip Trammell and Charles I. Jones argue that GDP is poorly suited to counting the things that improve our lives most: namely, new goods, higher-quality goods, and the improvements that never register as a price at all. They open with Nathan Rothschild, the richest man in the world in the 1830s, who died at 58 in 1836 of an infection that $10 of antibiotics could likely cure today.

Picture an economy of two goods — food, which we get better at producing every year, and string quartets, which take the same four musicians they always have. As food gets cheaper, people spend a growing share of their money on music — and since the growth rate is a reflection of spending, the rate drifts downwards. Push that to its conclusion and you reach the absurd result that we would have been richer had the string quartet never been invented. A slowing growth rate, then, can be the mark of an economy rich enough to afford the finer things.

Progress comes from invention, not accumulation — and invention is precisely what GDP is worst at counting. The Trammell and Jones alternative is to use the value of a statistical life (VSL) — namely, what people will pay to reduce their risk of dying by a fraction. That figure reflects how much they value being alive, so it sweeps up everything that makes life better, new goods included. From this you can infer how fast living standards are really rising. As James Pethokoukis explains:

“The authors get at VSL partly by looking at the extra pay workers require to take more dangerous jobs. If workers need $1,000 more per year to accept a job with one extra death per 10,000 workers, that implies a statistical-life value of about $10 million.”

On the authors’ baseline, American lifetime wellbeing has risen five- to sevenfold since 1940. The usual consumption-based measures suggest it has merely doubled.

As the authors point out, the answer moves a great deal depending on the interest rate you assume. On less generous assumptions the gains shrink, or vanish. As Eamonn Ives has argued here, GDP shouldn’t be dismissed entirely. But we should be awake to the risk that relying on it too heavily understates both how much innovation is worth to us and the cost of putting up barriers to it.

Share

🏛️ Ian Ng, Researcher

A report published by GovAI analysed 375 LLM releases over the past eight years and found that 11% were delayed or never released to the EU, with the figure standing at 7% for the UK. While the UK saw no delay in accessing the single most capable model at any point, it is still concerning when states are increasingly relying on frontier models for cyberdefence. Any delay in access, while hostile actors face no such constraint, could open a dangerous gap.

The authors attribute most delays and non-releases to regulatory factors, with data protection law the main barrier. Although the UK inherited most of its data regime from the EU, the authors point to the EU’s slower clarification of regulations and its more fragmented enforcement as the reason why barriers were higher across the Channel. They also note that the overall decline in delays may reflect labs building more mature compliance functions and greater regulatory certainty over time.

If Europe wants homegrown startups building frontier models, it needs to reduce the regulatory burden on them. The fewer resources a startup must dedicate to navigating regulatory uncertainty, the more it can spend on building. The regulations also risk becoming an entrance barrier for startups while protecting incumbents with large legal teams from competition.

We may also be witnessing the fading away of the Brussels Effect. The Brussels Effect held that EU regulations would set the global standard because it’s easier for businesses to comply with the strictest rule than to maintain a separate product. However, we are increasingly seeing the European market being carved out from the initial rollout of emerging tech products owing to regulatory uncertainty. As the authors have noted, Meta’s Threads and Apple Intelligence were both delayed in their European releases to comply with the Digital Markets Act.

When the term Brussels Effect was coined in 2012, the EU’s GDP was roughly a quarter of the world’s — today it stands at approximately one-sixth. Mounting regulatory uncertainty sits poorly with a shrinking market, and it is only making it more convenient for emerging tech companies to delay or completely avoid rolling out their products in Europe.

GovAI researchers found no evidence the EU AI Act itself has yet caused a single delay, but the enforcement of its general-purpose AI rules will only begin this August. The EU is about to run that experiment in real time. Britain’s advantage is that it doesn’t have to follow it down the same path.

Become a Supporter or Adviser

🎖️ Philip Salter, Founder

The King’s Birthday Honours landed last week, which gives me an excuse to return to ask the same question we’ve been asking since 2021: how many of these honours go to the people inventing and building things?

Five years ago, in our report Honours for Innovators, Ned Donovan and Anton Howes ran the numbers and found the answer was not many. We have now repeated the exercise across the last eight lists — every New Year and Birthday Honours since the end of 2022, more than 9,000 appointments to the Order of the British Empire — to see whether anything has shifted.

It hasn’t. Just one in ten citations mentions anything to do with innovation, science or industry, almost exactly where we found it in 2021. Strip out the broadest catch-alls of “industry” and “business” and it is lower still. Narrow it to the word “entrepreneurship” and it all but vanishes: 0.68% of honours. And these are generous counts — plenty of the citations we include are really for management or charitable work within a sector, rather than for invention in its own right.

As we argued in Honours for Innovators: “Given the current system, one would be forgiven for assuming that the surest way to an honour is to become a civil servant, politician, or philanthropist, or to achieve the fame that comes naturally to especially successful sports people, musicians, authors and actors.”

There are always honourable exceptions. In the latest list, our Patron Chris Hulatt, co-founder of Octopus, was made a CBE for services to entrepreneurship. More of this, please. Or, if the Government is feeling more ambitious, they might institute our dedicated order — what we called the Elizabethan Order — of genuinely equal standing to the OBE, with the same familiar four classes and a Sir or Dame at the top, awarded purely for invention and enterprise. It would cost around £66,000 a year: less than a single MP’s salary, for a payoff in status and aspiration many times larger.

This isn’t as radical as it might sound. Britain built its early reputation as the best place in the world to innovate partly by heaping status on inventors: the Society of Arts (now the RSA) struck medals to encourage them, monarchs granted them personal pensions, and there was even a chivalric order — the Royal Guelphic — that honoured the likes of Charles Babbage and William Herschel, the astronomer who discovered Uranus. It lapsed in 1837, not because the idea failed but because Queen Victoria could not inherit the crown of Hanover.

It’s time for a new chivalric order. This is the signal we need to give that Britain is serious about being the best place in the world to be a scientist, an inventor or a founder.

Subscribe now

🔀 Mann Virdee, Head of Science and Technology

On Monday, the Government Office for Science published their updated five AI scenarios for 2030. It’s the first update to a set of scenarios originally developed in 2023 and first published last year — with the aim of helping policymakers plan for the future of AI.

Before getting into the details of the five scenarios, there’s an important methodological point. People often believe that scenarios are attempts to predict the future — but that’s not true. Scenarios are a rigorous and methodical way to consider several imagined future situations which could come to pass, but they don’t have to happen in order to be useful. They’re designed to be different from one another, and their value lies in helping policymakers identify trends and useful courses of action across a wide range of potential outcomes. That’s to say, we should avoid the temptation to focus on the scenario we think is most likely because that’s not the purpose of such exercises. At the same time, they should still reflect plausible outcomes, otherwise you end up with AI scenarios like this unhelpful graph from the Federal Reserve Bank of Dallas.

The scenarios outlined in the GO-Science report range from a ‘slow burn’, to ‘augmented growth’, to ‘take off’ — and provide a more nuanced picture than the graph above. The report looks at six ‘critical uncertainties’ for each scenario: capability, model access, security, adoption, labour displacement and global cooperation.

Across most scenarios, AI drives significant productivity gains, helps to transform public services and make them more accessible, and accelerates scientific breakthroughs in fields such as health and energy — which will likely become key drivers of Britain’s productivity growth.

At the same time, across all scenarios, even in the slowest, the nature of cognitive work changes significantly, with routine, execution-oriented tasks being automated. There’s also a risk that workers become overly reliant on AI and have trouble operating if it fails. Another finding that holds true across all scenarios is the uneven adoption of AI and the compounding effect that will have, with a bifurcation where some realise the tremendous potential of AI while others are left behind.

It may not be the most groundbreaking report, but should serve as a good tool to help policymakers think more systematically about the future of AI and its adoption.

Share

⚡Ian Ng, Researcher

The Trump administration placed Anthropic’s Fable 5 under export control last Friday, barring foreign nationals from accessing it. Anthropic responded by disabling the model for all customers. The ban came a day after the Europe 2031 essay imagined precisely a world in which Washington rations AI exports as a geopolitical lever. AI Minister Kanishka Narayan drew the obvious lesson: “access to AI capabilities is crucial.”

Europe’s dilemma is rooted in having no frontier model of its own. Adopt American models and you cede control of the access lever; adopt them slower than the US and the productivity gap only widens. If the past year has demonstrated anything, it’s the value of sovereignty and autonomy.

We live in an economy defined by chokepoints. Just as the Netherlands has ASML and Taiwan TSMC, Britain must think about the leverage we can build. We do have a seat at the table as a signatory of Pax Silica and the AI Security Institute being one of the few trusted to evaluate Mythos. However, it does not guarantee access to frontier AI models. The US has already floated a “trusted partner” scheme granting close allies privileged access. The tiers are being drawn now, and Britain cannot be sure if it will retain such access in future.

Leverage cannot be built overnight. The nature of a chokepoint is that once built, it cannot be easily replicated. ASML took four decades and an extraordinary accumulation of tacit knowledge across 5,000 suppliers. But that is precisely the argument for doubling down now. Britain, through our higher education sector, still holds an edge on talent. Our universities produced the researchers behind DeepMind and Arm. Without strategic chokepoints of our own, it matters all the more that we sharpen the edge that we still have and do everything to retain that talent.

Keeping talent is not solely about money because Britain will not win a bidding war with American labs. It is about whether there is anything here to work on. For frontier researchers that means compute, and the gap is stark: Isambard-AI, our most powerful machine, ranks eleventh in the world, while the top three are all American exascale systems.

We are not going to out-spend Washington. Government funding for British compute — £1 billion for AIRR, £750 million for Edinburgh — is barely a tenth of the £22 billion Microsoft alone is putting into the UK. Accepting American capital is unavoidable, and any honest plan builds most of its capacity that way. But some must sit on a sovereign core — publicly held compute backing British firms directly. AIRR is how Britain avoids having to 'rent its AI future from abroad’.

If the shelving of the Edinburgh supercomputer and the pausing of OpenAI’s Stargate UK were not wake-up calls enough, the export ban on Fable 5 should be. Certainty over the long term funding is crucial to attracting investment and retaining talent. That requires fiscal discipline from the government — both resisting borrowing for day-to-day spending as well as resisting the urge to axe capital projects when money is needed elsewhere.

None of this delivers leverage Britain can wield alone. But the fundamentals built at home are what give us something to bring to a table we cannot dominate.

Become a Supporter or Adviser

🏗️ Philip Salter, Founder

Since 1970, productivity across most of the American economy has roughly doubled. In construction, it has fallen by around 40%. A new VoxEU column by Dongkeun Choi and Munseob Lee unpacks why.

The fall in the price of equipment — computers, machines, instruments — has been one of the great engines of the post-war economy, adding around 1.3 percentage points a year to growth in output per person. But structures have moved the other way. The relative price of buildings in the US is now 80% higher than in 1970, and that rise claws back almost two-fifths of the gain from cheaper machines. The net contribution of falling capital-goods prices is therefore closer to 0.8 points than 1.3.

“About three-quarters of the drag runs through standard capital deepening. When structures are expensive, firms accumulate less of them, and production slows accordingly. The remainder operates through innovation. Laboratories, offices, and pilot plants are themselves structures. Stagnant productivity in construction raises the cost of doing science.”

This is not an American curiosity. Choi and Lee examine thirteen advanced economies, and all but Belgium sit in the same troubling quadrant: construction prices up, construction productivity down. Across the entire sample, the UK records both the largest fall in construction productivity and the steepest rise in the relative price of building.

Why has construction forgotten how to build? The leading suspect is regulation. Hilber and Vermeulen show that the restrictiveness of the UK’s planning system, more than any physical shortage of land, drives the long-run rise in house prices; D’Amico and co-authors tie America’s construction-productivity stagnation directly to land-use rules. A planning regime that makes every project bespoke, contested and slow has meant construction is one of the few industries that never industrialised — it never achieved the scale economies and standardisation that lifted output almost everywhere else.

This resembles Baumol’s cost disease. When productivity stalls in one sector but the rest of the economy still needs its output, the relative price rises and everyone else pays for it. What makes construction unusual is that there is no way to route around it: the economy cannot make do with fewer hospitals, fewer fabs or — increasingly — fewer data centres. The cost of standing still in construction shows up everywhere.

As is often argued, restrictive planning acts as a tax on housebuilding. But it has also held back innovation in the construction industry. Alongside planning reform, we need to look deeper at what’s made us less efficient at building.

Subscribe now

🏹 Mann Virdee, Head of Science and Technology

When I was invited to give evidence before the Business and Trade Select Committee on industrial strategy, I emphasised three main points. First, a few outliers skew the statistics on British science. Once they’re removed, British science isn’t quite so ‘world-leading’. Second, the state can play an important role in procurement, such as through Advanced Market Commitments, and in de-risking the journey to market for entrepreneurs. Third, I offered some historical background on how Silicon Valley came to be the world’s pre-eminent hub for innovation and entrepreneurship.

But one question from the committee stumped me slightly: how effective is the Catapult Network? It’s a part of the UK’s R&D ecosystem I hadn’t really looked into in detail, although my overwhelming sense was that the Catapults were usually an afterthought in conversations about innovation and commercialisation I’d been part of. I thought it best to say nothing rather than pretending I had a more considered response.

The Catapult Network was created in 2011 after a report by Hermann Hauser that proposed an elite network of centres to help translate breakthrough scientific discoveries into commercial industries. It was modelled on 12 international comparators, including Germany’s Fraunhofer institutes.

There have been a series of reviews with mixed findings. A 2014 review called for doubling down on the approach, saying that it was mirroring international comparators, and recommended expansion. A 2017 review by Ernst & Young found that the centres were not being properly managed and that they had no common purpose statement. A 2021 government review recommended reviewing the Catapults less often, but it also found that the High Value Manufacturing Catapult alone had generated 75% of all the Catapults’ income the previous year, showing a highly uneven impact.

Against this background, there are reports that ministers are lining up another review of the Catapults to assess their value and impact after concerns that some have failed to support regional growth and help build national champions. It’s rumoured that streamlining and job cuts may be on the cards.

I recently wrote about an OECD report on the ‘valley of death’ between Britain’s strong support for research up to prototype and its thin support for demonstration, customer validation and early market entry. That report’s proposed solution was to expand the commercialisation role of the Catapults.

So the function clearly matters. The gap the Catapults were built to fill is, if anything, widening. The question remains whether these particular institutions are still the right vehicle for the job.

One approach is to keep tinkering and topping up funding, hoping that some future permutation works. The other is to know when to call it a day and build something new with a sharper remit, explicitly tied to growth and closing the demonstration-to-market gap. Founders I’ve spoken to lean towards the latter. But before we can choose well between these, we need an honest diagnosis of why the Catapults are underperforming — and what, concretely, we would do differently.

Share

📈 Rafi Pollack-Joyce, Policy Analyst, Public First

Tony Blair’s intervention last week has put AI in the public sector at the heart of the fledgling Labour leadership debate. But is he right that governments can harness the technology to deliver more with less?

Earlier this year, Public First surveyed 3,335 public sector workers across ten countries. The headline finding is striking. AI is everywhere. Around three-quarters of public servants now use it, and most started in the past year. That probably makes AI the fastest-adopted technology the public sector has ever seen. But there’s a big difference between using a tool and changing how the government works.

The countries doing best aren’t simply the ones with the biggest AI sectors. They’re the ones that have made AI feel usable inside government. That means clear permission, decent training, approved tools, and a way for good experiments to become normal practice.

Singapore is the clearest example. Its advantage isn’t magic technology. It’s that public servants have more of the scaffolding around them: guidance, tools, training and institutional support. For example, Singapore is twice as likely as the UK or US to conduct mandatory training for employees. The results are clear: compared with the UK and US, Singaporean public sector workers are more than twice as likely to use AI daily, to be using it for complex tasks, and to think the public sector in their country overall is using it effectively.

The UK and US have a more awkward problem. Both are AI leaders in the obvious sense, with companies, researchers and policy attention. But inside government, use is patchier. People are interested, and often already experimenting, but many still don’t have clear guidance on what they’re allowed to do or how to move beyond low-risk tasks. While just over half of public servants in the UK and US feel confident using AI tools, that rises to 85% in Singapore.

That matters because unclear permission doesn’t necessarily stop AI use. It just makes it messier. People experiment on personal accounts, stick to shallow use cases, or run pilots that never really scale.

Ultimately, this is fixable. The hard part isn’t persuading public servants that AI matters, it’s building the basic machinery around it: procurement, guidance, training, data access and routes to scale.

Blair is right that AI could change the state. But the first test is more mundane: whether the government can manage the adoption that’s already happening.

Become a Supporter or Adviser

When the electric motor was first invented, it didn’t immediately improve productivity in factories that had previously relied on steam engines. As steam engines were so large and their power distributed mechanically, their setup determined how electric motors were integrated into workflows. Electric motors were initially used to replace a central large source of power — but it was later discovered that the efficiency gains and benefits really started when that central source was replaced by a larger number of smaller motors.

Such is often the case in innovation. We might have the means of improving processes but it takes time for them to be properly adopted and adapted to their intended uses. This is particularly true in Britain’s £747 billion engineering industry, where breaking out of the conventional way of operating can be difficult.

Engineers currently design individual components as best they can, and then piece them together into a complex system later. But that approach is facing challenges driven in part by resource scarcity, but also a desire to achieve greater efficiency. The solution may be to reimagine our whole approach to design by thinking of complex hardware as a single, holistic entity that can be designed and optimised at the same time.

Jarek Rzepecki, the new CEO of Monumo, believes this is the foundation for all engineering in the coming decades. By moving away from component-led design toward a system-level optimisation, Monumo is showing that significant cost reductions can be made without sacrificing performance.

If Monumo succeeds, it will enable a leap in industrial productivity without the energy and infrastructure requirements of traditional AI models. I sat down with Jarek to discuss his transition from CTO to CEO, the systemic challenges facing the UK’s deep-tech investment landscape, and how Monumo is leveraging foundational physics and its Anser AI computation engine to radically optimise the design and sustainability of electric motors.

What we discussed

• Jarek’s transition from CTO to CEO and how to balance technical rigor with investor and customer relations.

• The shift from component-based engineering to holistic, AI-driven system design.

• Why the UK’s focus should move from data centres for large language models towards the intersection of AI, physics and chemistry.

• The structural opportunity for AI-native manufacturing in the UK.

• Why the path to UK productivity lies in building new industries from scratch rather than retrofitting incumbents.

Lessons for scaling deep-tech hardware companies

For policymakers:

• Prioritise ‘AI for Science’ over LLMs. The UK doesn’t have a comparative advantage for LLMs considering the energy and infrastructure required. Instead, the UK should focus on the intersection of AI, physics and chemistry, where the UK has a distinct research advantage and lower infrastructure costs.

• Build AI-native manufacturing. Rather than trying to turn around incumbent industries, the UK should create new, AI-native industries that are built on the latest techniques from day one.

• Lighten the grant admin burden. Set a high bar to win grants, but reduce the ongoing reporting and administrative load once awarded — including for any new schemes. The current burden falls heaviest on the small companies grants are meant to support.

• Fix the London listing environment. The current cost-benefit analysis makes it unwise to list in London. Improving the appeal of the London Stock Exchange would provide a benefit to the economy that far outweighs any loss in immediate oversight or the administrative costs of transition.

• Fund fundamental AI research. UK grant structures currently require AI research to be tied to manufacturing applications, making it nearly impossible to fund pure AI research.

For founders:

• Prepare for disruption. Deeply disruptive technologies face a credibility gap. Founders must be prepared to build trust and credibility over time as incumbents get comfortable with new, AI-native ways of doing business.

• Focus on resilience. Building a company is never easy. Beyond technical curiosity, the tenacity and creativity to build what hasn’t existed before are the two primary qualities of a successful entrepreneur.

• Own the full stack. While this is not possible for many businesses, for founders in the deep-tech or AI space, building your own foundational models rooted in first principles rather than building on top of someone else’s API can offer tremendous advantages.

• Prove yourself in one market, then diversify. Cross the Valley of Death with radical focus before expanding. Once your core technology is proven, diversification across sectors becomes a strength — especially when the underlying technology is horizontally applicable.

Full interview

I. Building Monumo

Monumo applies foundational physics to hardware architecture. Can you explain what you’re building and why it matters?

Monumo is an engineering AI company focused on creating complex engineering designs, at speed and ready for production. We are currently working with global brands in the automotive, domestic appliance and industrial sectors, initially focused on electric motor systems, with robotics and aviation next on the roadmap.

Monumo’s Anser® Engine uses deep-tech and AI to explore millions of potential design options, evaluates them against set customer parameters and objectives and selects the optimal designs for production in just hours.

The Anser Engine does not require customer data to create potential designs but just the requirements it must fulfil, such as shape, size, torque, efficiency or cost. After understanding the requirements, the Anser Engine generates its own data based on high-precision simulations and then assesses each design for fit. It generates results at incredible speeds reaching beyond human capabilities. So what we’re building is the foundation for how all engineering will be done in the coming decades, such as for cars, planes, industrial equipment and energy turbines.

Can you give me some examples of projects you’ve worked on?

In one recent project, the Anser Engine revealed a 17% cost reduction over a customer’s original electric motor design, enabling savings of up to €70 per motor.

In terms of how that 17% cost reduction was achieved, in an electric motor, there are three main materials: electrical steel for the stator and rotor, copper for the windings, and permanent magnets. Of all these materials, the magnets are by far the most expensive. In this particular project, we managed to reduce the required volume of permanent magnets. By moving the magnets around, changing the shapes of air pockets, and adjusting the design, we achieved a motor with the same performance as the original but with a significantly lower material cost. It also helps reduce dependency on the volatile rare-earth supply chain.

When you consider that the top 10 automotive OEMs [Original Equipment Manufacturers] will need many new motor designs in the next decade, the potential for industry-wide savings is staggering.

Aside from electric motors, what else are you looking to optimise?

At the moment, we are working exclusively with electric motors, but across very different sectors. A motor for a washing machine is quite different from one for an EV or a drone. They use the same underlying physics, which is why our technology works for all of them, but the metrics for success differ.

In drones, weight is the priority. Every gram saved is a gram the drone doesn’t have to lift. For industrial applications, efficiency is key. These motors often run 24/7, so even small electricity savings add up to huge cost reductions. In the EV space, cost is the main driver to help manufacturers compete. In humanoid robotics, there is a strong need for dedicated actuators. Currently, companies either buy expensive high-quality actuators or hobbyist-grade ones that aren’t good enough. Monumo can produce designs for dedicated actuators that fit the specific requirements of humanoid robots.

Our vision is to be the go-to designer across all these verticals. Diversification is important because sectors have golden eras. Eight years ago, EVs were moving extremely fast; now, that has slowed down while robotics is kicking off. If you put all your eggs in one basket, you risk ending up in trouble. Since our technology applies across the board, doubling down on just one sector would mean missing out on revenue streams in white goods or industrial motors. Addressing as many markets as possible is the right choice for risk reduction.

In the distant future, we would like to be optimising the entirety of engineering. But optimising a whole car is orders of magnitude more complex than optimising the powertrain.

Our expansion will be gradual. If you take the drone example, we currently optimise the motor, controller, and inverter together. The next step could be optimising the shape of the propeller. We might not optimise the frame yet, but we’ll take it step-by-step. If we see a benefit at one stage, we move to the next. While a full drone is within our reach, we probably won’t be optimising car seatbelts or entertainment systems anytime soon!

What has enabled this transition from designing components separately to engineering the system as a whole in the last few years?

The big shift is around the speed of compute and the scalability of our Anser Engine. It’s also about having the right focus. In one recent project, we evaluated 3,000,000 designs in just three days. That level of iteration is simply impossible for human engineers using traditional methods.

If you keep doing that over the next few decades, we’ll end up being able to design an entire aircraft or car as a single entity. That’s really exciting because it will transform the performance and use of materials in so many products that are essential to modern society.

To power that kind of transformation, I’d imagine you can’t just rely on standard off-the-shelf LLMs. Do you train your own AI models, or use off-the-shelf ones?

For what we do, there are no off-the-shelf models, though we do use known architectures. Models have to be trained by us, from scratch, and we own the full stack. One thing to mention about AI for Science or Engineering is that it often does not require as much compute or the energy that frontier LLMs do. I really don’t think it’s the best use of resources to chase after the frontier LLMs in the UK, because we don’t really have a comparative advantage when you consider the energy and infrastructure required. We can get ahead in AI for Science or Engineering without those same requirements, especially if you leverage deep scientific and engineering expertise the UK does have.

You recently transitioned from CTO to CEO. Now that you’re stepping into the lead role, what different skill sets do you think this will take?

Monumo is a deep-tech company, so a huge portion of running the company was always centred around the technology. The development team is the core of the company. To some degree, I’ve been doing a lot of this since the beginning. I have also been heavily involved in VC conversations and customer interactions in the past.

In terms of skill set, I don’t feel there is a huge gap between what I have been doing and what I will need to do going forward. However, the ratios will change. In the past, I might spend 30% of my time with customers, 60% with technology, and 10% with VCs. Now, I imagine it will be a more even split of a third on each of these. I still want to stay close to the technology — it’s vital that I remain close to our technical developments — but being close to customers and VCs is equally important as we look to raise more funds. So I’d say the skills remain similar, only the focus shifts.

Tell me a bit more about Monumo’s history. How did it begin, what’s your team like, and has anyone else attempted to build something similar?

The idea came from research programmes at Arm. The founding team noticed that electric motors were underdeveloped and saw scope for improvement.

In terms of the team itself, we have quite a few Arm veterans in the business. Over half of the team have PhDs in physics, electronics, machine learning, and computer science. It’s a very science-based team in Cambridge, and then we have a team in Coventry that has the domain expertise in automotive engineering and prototyping.

Regarding competitors, there are three main groups, but no one does everything we do. There are incumbent simulation companies such as Ansys and Siemens. They have high-fidelity simulators, but their approach is very human-centric. Their stacks weren’t built from the ground up for AI. Then there are AI-for-Physics startups, which are trying to crack AI for engineering but they often lack access to high-quality simulation or in-house engineering expertise. There are also engineering giants who have huge domain expertise but lack the specialised simulation or AI expertise.

Monumo sits in the middle of those three. We combine domain expertise via our hardware team in Coventry, our own high-quality simulator, and an AI team that uses that data to train models. To my knowledge, we are the only ones with all three pieces.

II. Navigating the UK’s industrial investment landscape

What has been your experience with the UK’s funding landscape?

R&D tax credits have been enormously helpful, and I hope they don’t change. But the grant process is quite onerous. I think there should be a high bar for getting grants in the first place, but once the grant has been made, the administration and reporting should be less onerous. At the moment, they put a huge, continuing burden on businesses.

One of the issues is that emerging technologies such as AI are so deeply disruptive to long-established companies that there is a natural timidness and lag around adopting them. Support to help large companies start the transition and the evolution towards an AI-native way of business would be helpful.

You have spoken about the UK’s exit problem. What needs to change?

At the moment, I don’t think I could find a single adviser who would say that if we were to float the company, we should do it in London. It comes down to a fundamental valuation gap. If you look at multiples paid on the Nasdaq versus the London Stock Exchange (LSE), a deep-tech company often sees a 30–50% discount just for being listed in the UK. That’s largely because we lack a dense ecosystem of institutional investors and analysts who truly understand deep-tech timelines. Most British funds are looking for immediate dividends or steady cash flow, whereas a company like Monumo requires a decade-long view.

Liquidity is also a massive issue. There aren’t enough anchor tech investors in London to keep the stock price stable during the early years of a listing. For a founder, if you list in London you just aren’t getting the visibility or the capital you need to compete with a Silicon Valley peer.

When a British startup reaches a certain scale, they often face a choice: selling to an American or Chinese giant, or doing an IPO in London. From a government perspective, a trade sale provides a quick, one-time shift in revenue via capital gains tax and a clean exit for early investors. But the country loses the IP, the high-paying head-office jobs, and the future tax base of a trillion-dollar company. I’m arguing that we should prioritise the long-term health of the LSE over those short-term wins.

III. Conclusion and advice

If you were advising the Secretary of State for Science, Innovation and Technology, what would be your priorities?

It would be good if the government talked a little less and did a little more. There is a lot of talk about AI sovereignty, but that needs to translate into real help for tech companies. It’s still significantly easier to access funds in the US than in Europe.

One specific issue is government grants. It is very hard to find grants that support pure AI research; they usually require you to link it to manufacturing. Fundamental research is the bottom of everything. Companies like OpenAI or Anthropic are software companies with heavy foundational AI components; the UK doesn’t really have an equivalent domestic AI powerhouse. This is a strategic problem. AI is a strategic resource, and relying on other countries for it is dangerous.

What is your advice to founders in this space?

I wish this wasn’t the case, but you have to be extremely resilient to cope with building a company. It’s never easy. You also need more creativity than is often thought — you really need to be able to build something that hasn’t existed before.

You have to be prepared to cross the Valley of Death into more fertile territory. You also need to maintain a radical focus, prove yourself in one market first before you get distracted by the multiple potential applications of a deep-tech solution.

We ask all our guests the same closing question: what’s one interesting thing you’ve read or listened to recently that you’d like to share?

Recently, I read The Magic Machine: A Handbook of Computer Sorcery by A. K. Dewdney. It’s a delightful playground for technical minds who want to rediscover the joy of computing. It covers a surprisingly broad range of topics from chaos theory to early graphics, all paired with hands-on programming challenges. The author has a rare gift for making deep computational ideas feel like puzzles you actually want to solve. For me, it was a trip down memory lane, revisiting how computer programming was in the days when I was growing up and learning it for the first time.

🧑‍💼 Philip Salter, Founder

AI adoption is usually discussed in terms of cost and compute. In More Than Just Plug and Play, a working paper published this month, Diane Coyle and colleagues at the Bennett School use the ONS Management and Expectations Survey to show that UK firms with stronger management practices in 2020 were significantly more likely to go on to adopt AI by 2023.

Digging a bit deeper, it turns out different technologies call for different organisational capabilities. The same management practices that predict AI adoption show no relationship with the adoption of robotics, specialised software or specialised equipment.

We already know the UK has a management problem. Nick Bloom and John Van Reenen’s seminal 2007 paper introduced a structured way of measuring management practices across firms and countries, and the body of work that followed has consistently shown UK firms lagging US counterparts — with a particularly long tail of badly managed firms dragging down the productivity distribution. The policy response has tended to bundle this into general management improvement — training, peer learning and leadership development — rather than targeting the specific practices.

The new evidence narrows the target. As the paper notes, “it is specifically monitoring practices, such as use of KPIs and target-tracking, along with decentralised product development, that predict adoption.” Practices around continuous improvement and employment had weaker or insignificant effects.

Among multi-site firms, those where product development decisions sit closer to the frontline were more likely to adopt AI. The authors suggest “AI applications are more likely to be context-specific and require domain or technical knowledge to identify valuable use cases.” The combination of autonomy and measurement is the specific organisational structure that predicts AI adoption.

The policy implication is that management training interventions may be better targeted on capability-building around performance measurement and data architecture, combined with organisational design that empowers technical teams.

Policy aside, there’s a lesson here for founders bullish on AI. You don’t need a government programme to adopt the organisational shape this research describes. Building KPI infrastructure and pushing product decisions closer to the people who understand the work is something firms can do on their own. The companies that get this right won’t be waiting for policy to catch up. (Mann picks up the structural side of this problem below.)

Subscribe now

💻 Mann Virdee, Head of Science and Technology

The OECD, in collaboration with the Department for Science, Innovation and Technology, recently released a report on technology adoption in the UK.

On the one hand, their findings indicate that British firms do well on mature digital technologies. The UK’s adoption of cloud computing, data analytics and basic process tools sits above EU and OECD averages, and SMEs have largely closed the gap with larger firms in using these foundational tools. On the other hand, for more advanced technologies, such as AI, robotics and automation, uptake is more limited than one would expect for a country with the UK’s income level (see Philip’s article above).

Take robotics, for example. The UK has a strong manufacturing legacy, particularly in the Midlands — yet adoption of robotics among manufacturing firms trails the EU average. The OECD attributes this to a mix of factors that compound in smaller manufacturers: high upfront costs, the average age of business owners, distrust of technology vendors, and a ‘wait-and-see’ approach whereby firms only adopt after peer validation. Each barrier is manageable in isolation, but together they create a structural drag on uptake in the sector in which the productivity dividend should be largest.

Geography compounds this problem. The OECD’s data on regional adoption finds that between 2018 and 2022, average SME adoption of AI and robotics technologies stood at around 15% in London and the South-East, but just 1.9% in the North-East. The concentration in the Golden Triangle is well-documented, but the within-region picture is just as stark. In the West Midlands, the Black Country trails most innovation indicators despite sitting next to one of the UK’s strongest industrial clusters.

The OECD’s overarching diagnosis is that Britain supports research well through to prototype (Technology Readiness Level (TRL) 6–7), but that support thins out at the stage of demonstration at scale, customer validation and early market entry (TRL 8–9). It’s at the later stage that firms package innovations into off-the-shelf products with the reliability and support that allow non-frontier SMEs to adopt them. Without that later stage of support, brilliant science struggles to become widely diffused technology.

The report’s proposed solution is to expand the commercialisation role of the Catapults, paired with the British Business Bank to crowd in private co-investment. For our part at The Entrepreneurs Network, we’ve been bringing together robotics founders with the Regulatory Innovation Office and the Health and Safety Executive to try to make their scaling journey smoother — exactly the kind of intermediary work the TRL 8–9 gap demands at scale.

Share

🛡️ Harry Pitts, University of Exeter

Britain’s drive towards defence reindustrialisation and rearmament will depend on whether smaller, more agile firms can be integrated into the defence supply chain. In a new report for Babcock, The Next Line of Defence: Unlocking SME Potential in UK Defence from Policy to Practice, our team at the University of Exeter’s Defence, Security & Resilience Network found that the primary barriers relate to culture and process.

We interviewed 20 cutting-edge defence SMEs, both those well-established in the supply chain and those pivoting in from civilian sectors. The picture that emerged is of significant capability waiting to be unlocked, set against procurement systems that struggle to engage with it.

Ukraine’s high-tech resistance to Russia’s invasion has demonstrated what agile, software-led innovation can do on the modern battlefield. The UK has firms capable of producing similar capabilities at pace, but defence procurement, administrative systems and financing frameworks remain designed around large, established suppliers operating on multi-year cycles.

SMEs need credible demand signals to justify the investment and recruitment that scaling for defence requires. The forthcoming Defence Investment Plan is the obvious vehicle, and its credibility with the SME community will depend on how concretely it signals where money will flow and over what timeframe.

There is also a vital role for primes — the large, established contractors that sit between government and the SME base. Our interviews suggest that the most productive prime–SME relationships invert the usual power dynamic, with primes acting as supporting partners that help smaller companies articulate the value of their products to government, navigate procurement processes and bridge the gap between SME agility and government risk aversion. This is a markedly different role from the traditional one of subcontracting work down a tiered supply chain.

Babcock’s SME Engagement Charter, informed by our evidence, is an attempt to formalise this. Against a backdrop of rapid geopolitical change, primes have the power to realise in practice the promise of government policy in this domain.

Become a Supporter or Adviser

Subscribe now

🧠 Mann Virdee, Head of Science and Technology

Google DeepMind is perhaps the most transformative startup Britain has produced. It’s now synonymous in the minds of many, myself included, with its protein structure prediction system, AlphaFold, and the 2024 Nobel Prize in Chemistry.

But did you know that to secure early funding, DeepMind’s CEO and co-founder Demis Hassabis once had to pitch investor David Gammon at his home — during which he was also required to win approval from Gammon’s wife and three teenage sons? It seems strange now that such a successful company could have been set back or derailed if just one of them hadn’t been convinced by his pitch.

It’s one of many stories in Sebastian Mallaby’s new book The Infinity Machine, which follows DeepMind’s rise from a startup in a Russell Square townhouse to the heart of Google’s operations. It’s well worth a read for those who want to understand what it takes to be a successful entrepreneur, as well as to explore broader questions about the future of scientific research, the nature of consciousness and the potential of artificial intelligence.

The family-veto story is important. In 2010, no British venture capitalist was set up to fund a company whose stated mission was to ‘solve intelligence’. That meant that Hassabis had to piece together funding from Peter Thiel, Elon Musk and others. By the start of 2014, the money had run out and the only viable option was to sell to Google.

Some mistakenly draw the conclusion DeepMind should not have sold to Google. But Google’s compute and resources were necessary for DeepMind to thrive. The lesson is a broader one. Britain still struggles to fund moonshots at the scale and time horizon they need.

It’s been said time and time again that our universities are world class. So too are our founders. But the patient, multi-billion-pound bets that turn blue-sky research into world-leading companies remain largely an American endowment. Closing that gap is partly about R&D and tax design, and partly about cultivating high-conviction angels and family offices willing to back unproven founders with audacious ideas.

That makes the recent news from David Silver, Hassabis’s old Cambridge friend and the man behind AlphaGo, feel like a test. Silver’s London startup Ineffable Intelligence has raised $1.1 billion at a $5.1 billion valuation, which is the largest seed round in European history, with the UK’s Sovereign AI Fund investing alongside Sequoia, Lightspeed, Nvidia and Google. The capital is still mostly from the US, but the company is British. It remains to be seen whether we can keep it that way as Ineffable scales — and whether we can grow and support more companies like it.

Share

🐝 Philip Salter, Founder

Britain’s productivity problem is, in part, a problem of its big cities outside London underperforming. Manchester, Birmingham, Leeds, Glasgow — in most comparable countries, second-tier cities punch above the national average. In ours, they don’t. Closing that gap matters not just for the people who live there but for national prosperity.

A new CEP paper from Aadya Bahl and Henry Overman, Hive of Talent, uses Greater Manchester as a case study to ask what it would actually take to close one component of that gap: skills. The numbers are sobering. Manchester’s productivity sits 35% below London’s, against a 20% gap between, for example, Paris and Lyon. Even partly closing it would require — among other things — an additional 180,000 workers with degree or sub-degree qualifications.

The paper considers three pathways through which Manchester adds graduates: people who grow up there, people who move there to study, and people who move there for work. On the local cohort, even matching London on both attainment and retention, the GCSE pathway would generate only an extra 4,285 graduates a year. The local cohort is “just too small relative to the size of the workforce for even quite sizeable improvements in attainment or retention to generate the scale of changes needed.”

Manchester’s universities attract and retain around 11,400 graduates from outside the city region each year, but a meaningful slice of that depends on international students. The Government’s Immigration White Paper may dampen this significantly. And on non-education migration, the picture is similarly fragile. Manchester currently gains around 2,606 graduates a year through net international migration but loses 1,219 to the rest of the UK, for a net gain of about 1,387. Halve net international migration and the net gain shrinks to just 84 a year — a fall of roughly 1,300 graduates through this channel alone.

As the paper acknowledges, keeping more skilled workers in Manchester only helps if there are skilled jobs there for them to do:

“London benefits from a unique concentration of universities, high-skilled employment opportunities, and strong graduate labour markets that attract and retain graduates from across the UK and internationally… replicating these conditions in [Greater Manchester] would require changes that go well beyond skills policy alone, as factors such as labour market opportunities, wages, and supporting infrastructure all shape graduate location choices.”

The wider lesson is that no single lever moves the dial, and the supply-side levers don’t move it at all if the demand-side jobs aren’t there to absorb the workers. For founders and policymakers tempted by simple stories about either just fixing schools or visas — important as both are — this is a useful corrective. Manchester’s skills problem is a coordination problem, and one the rest of Britain’s big cities, with smaller cohorts and weaker universities, will face in even less forgiving forms.

Become a Supporter or Adviser

💼 David Bharier, British Chambers of Commerce

By the late 1990s, most businesses had a website. While many firms simply replicated their brochures online, a smaller group — from Amazon to early digital-native retailers — built their operations around the technology entirely. It was that depth of adoption, not the speed of uptake, that ultimately separated the winners from the rest.

We may be seeing a similar pattern emerge with AI. Much of the current debate focuses on how quickly firms are adopting the technology. On that measure, the UK appears to be moving at pace. Analysis conducted with the University of Essex, using British Chambers of Commerce survey data from early 2026, finds that over half of SMEs are now using AI in some form, up sharply from around a third last year.

But this headline figure masks a more important reality: not all AI adoption is equal. For most firms, AI is not yet translating into meaningful workforce change. Around 95% of users report no impact on headcount, while 86% say job roles remain unchanged. This reflects the fact that most businesses are still using relatively light-touch, off-the-shelf tools — such as ChatGPT or Copilot — to support existing tasks rather than fundamentally redesign them.

The picture looks very different among the smaller group of firms embedding AI more deeply into their operations. Among these businesses, around one in five report staffing reductions attributable to AI, and they are more likely to have reorganised job roles. So it is not adoption itself that is driving change in the labour market, but the intensity of it. This distinction matters. Workforce restructuring may be concentrated among a relatively small but potentially growing segment of firms with adoption.

The key question now is whether this group remains niche or begins to grow. If it does, the lesson from the early internet era is clear: it will not be the firms that simply “use AI” that shape the future, but those that reorganise themselves around it.

In a labour market where wage floors have risen sharply and new firms are taking on fewer staff, the challenge for policymakers and businesses alike will be to ensure that this transition delivers productivity gains while bringing the workforce with it.

Subscribe now

🔬 Mann Virdee, Head of Science and Technology

It’s not new or insightful to note that Europe struggles when it comes to converting bright ideas and cutting-edge science into unicorns.

What is new, however, is a recently published comparative study in the Journal of European and International IP Law which seeks to explain why. The researchers looked at the legal frameworks of the United States, the United Kingdom and Italy and found a structural difference in how these countries deal with technology transfer (TT). In particular, they claim that the strictness of regulation is less important than the clarity of the regulation.

On one side you have the US model, rooted in the 1980 Bayh-Dole Act. This Act transformed TT by allowing universities, small businesses and non-profits to retain IP for inventions developed from federally funded research, as well as to license discoveries from that research to private-sector partners. That provided a big incentive for universities and researchers to turn discoveries into viable consumer products.

The authors of this study argue that the US model has succeeded because it has created a predictable and coherent regime which gave universities clear ownership of IP, and which also gave investors confidence to invest. The UK has tried to mirror the US model and has been moderately successful, resulting in what the authors call a ‘relatively mature environment’. Both the US and UK models benefit from having clear institutional ownership and Technology Transfer Offices (TTO).

On the other side, there’s Italy’s fragmented model. Italy produces lots of spin-outs but no unicorns (with the notable exception of Bending Spoons, a Milan-based tech conglomerate whose focus is on acquiring and managing products such as Evernote). In Italy, the lack of clear, universal legislation creates a network of ‘autonomous rules’ that leaves founders and investors in a state of constant uncertainty.

So while the US and UK have stricter regulations around academic-entrepreneurial roles than Italy, the authors of this research contend that the ‘coherence, accessibility, and institutional robustness of TT regulation are decisive factors for the emergence of university-born scale-ups’.

Or, to put it another way, clear regulations — even if stricter — are better than vague ones. When a founder knows exactly how many days they can consult (for example, around 13 days at Stanford) and who owns the IP, they can get on with the business of scaling.

Although the UK is in a strong position, there’s more work to be done to streamline TTO processes. Scientific excellence is only half the battle. The other half is ensuring our legal and institutional frameworks are configured to let that excellence evolve into global success.

Share

📚Philip Salter, Founder

In 2008, Clayton Christensen’s Disrupting Class was published, arguing that schools were on the verge of being disrupted. Built around a factory model of standardised content, uniform pacing and batch processing of children, he argued that they were structurally incapable of personalising learning to individual needs. Technology, he contended, offered a way out — not by improving the existing model, but by routing around it entirely.

Christensen predicted that 50% of high school courses would be delivered online by 2019. They weren’t. Billions had been spent putting computers into schools, yet the technology has simply been co-opted into the existing model of instruction.

Last week’s government announcement, which invites EdTech companies to build AI tutoring tools for disadvantaged pupils in the UK, is encouraging. As the press release acknowledges, private tutoring can accelerate learning by up to five months, but it remains the preserve of those whose parents can afford it.

The ambition is right, but Christensen’s insights call into question the design: “Plugging a disruptive innovation into an existing business model never results in transformation of the model; instead, the existing model co-opts the innovation to sustain how it operates.” The Pioneer Group — eight companies, curriculum-aligned, teacher co-designed and DfE safety-approved — is structurally set up to do exactly that. Given the constraints any government programme operates under, it’s perhaps inevitable.

The Pioneer Group will co-design, pilot, evaluate and report, with national rollout targeted for 2027. While this is a relatively ambitious timeline for government, the technology is moving at breakneck speed. There is a real risk this programme institutionalises a version of AI the market has already moved two generations beyond.

In truth, the disruption Christensen predicted is already happening. A motivated 14-year-old with a capable AI can already access something closer to a genuine Socratic dialogue on any subject than most classrooms offer — without a Pioneer Group, without curriculum alignment and without waiting until 2027.

Become a Supporter or Adviser

🎓Ayushma Maharjan, Centre for Policy Studies

The UK is one of the best places in the world to produce ideas. But few of those ideas are turned into commercial strength at home. Britain’s academic strength is hard to dispute. Ten of the world’s top 50 research universities are British. However, that scientific excellence has failed to proportionally translate into a dense network of R&D-active firms and the scale-up ecosystem needed to translate research into economic output.

The gap is clear in the data. Cambridge and Oxford rank only 69th and 77th globally as innovation clusters, a striking contrast to their top five standing in university rankings. A country can be brilliant at science and still underperform if little R&D happens inside businesses.

Analysis done by the Centre for Policy Studies suggests that the UK higher education sector performs a relatively high share of national R&D, at around 24%, compared with 11% in the United States. The UK business sector, by contrast, performs about 70% of total R&D, below the roughly 80% share seen in the United States. In leading innovation economies, every dollar invested in higher education R&D is matched by $7 to $9 in business R&D. In the UK, that ratio is less than $3.

One plausible explanation for the UK’s position is that Britain’s wider business environment is not attractive enough for firms to build, test and scale. This is clear from companies like AstraZeneca and OpenAI, which have chosen to halt investment or move activity elsewhere citing tax burden, regulatory complexity and high energy costs.

If Britain does not address this problem, the economic returns on British science will continue to be captured elsewhere. Evidence suggests 80% of UK university spin-out IPOs have taken place overseas since 2012. Likewise, despite ranking highly in AI research, the UK retains only 48% of its talent.

No amount of tax credits or industrial strategies can compensate for a hostile business environment. The UK needs a more innovation-friendly economy. For entrepreneurs, that means a simpler and predictable business environment, competitive taxation, and stronger incentives to stay and scale in Britain.

Nobody disputes that artificial intelligence poses considerable disruption to the labour market. Yet much of the commentary to date has focused on which jobs are on the road to obsolescence, and how workers — or those about to enter the workforce for the first time — should be preparing themselves for this new economic epoch. Alfie Pearce-Higgins, however, has a more interesting perspective on what is unfolding. The really important problem to resolve, he says, is not so much which jobs AI may or may not automate, but rather how the very structure of employment itself could change.

Alfie is the Co-Founder of Rodeo, a company that started out giving gig workers visibility over their own earnings. Following a legal dispute with a major delivery platform after reverse-engineering the algorithms that set gig workers’ pay, Rodeo pivoted and built a suite of AI-driven tools to help people take control of their careers and find the professional pathway they want to be on.

I recently sat down with Alfie to learn more about how he thinks AI is upending the labour market, what consequences this will have for government, and what needs to change because of it.

What we discussed

• Why more attention should be paid to how AI will reshape how work is contracted rather than how it will influence demand for specific jobs.

• The three converging forces — regulatory costs, legal classification and AI itself — that are driving a potential surge in gig work.

• Why the UK’s self-employed population is almost invisible to government, and what consequences this has for tax collection, pensions and welfare.

• Why the skills the education system should be teaching for the future are essentially impossible to predict right now.

• How AI has broken recruitment signalling, turning hiring into an arms race between automated applications and automated screening.

• The asymmetry of information in algorithmic pay, and why transparency — not abolition — is the right policy answer.

Share

Lessons for policymakers:

• Employment classification is urgently overdue for reform. The legal framework governing gig work was designed for the self-employed sole trader negotiating with individual clients — not for someone being algorithmically dispatched to jobs through an app.

• Government is making employment more expensive at exactly the wrong moment. The combination of rising National Insurance contributions and National Minimum Wage rates, plus the Employment Rights Act — none of which apply to gig workers — creates a powerful incentive to shift towards self-employment.

• The gig economy is a fiscal blind spot. Around 4 million people earn income through digital platforms in the UK. HM Revenue and Customs (HMRC) has had earnings data on many of them for two years but appears to have done little, if anything, useful with it.

• Deducting tax at source is more powerful than most people appreciate. A shift from payroll employment to self-employment does not just reduce headline tax rates — it dramatically reduces compliance. Government should think carefully about how to extend source-based collection into new areas of work.

• The retirement savings crisis in the gig economy is real and growing. The employer pension contribution model, which does much of the heavy lifting in getting people to save, does not reach gig workers.

• Youth unemployment needs AI-native solutions. The traditional model of a face-to-face meeting to assess a young person’s situation is unscalable, and the labour market these people face is more opaque than it has ever been.

• Data portability rights for gig workers should be extended. Gig workers should have the same rights to access and use their earnings data as employees do. The current situation, in which platforms can contractually prohibit disclosure of pay, is an anomaly that policymakers should close.

Full interview

Before we properly dive into things, I think it’s important to first hear about the journey you’ve been on to get to where you are now, and, second, to get a sense of what you’re building now and your plans for the future. Can you shed some light on those for us?

Before co-founding Rodeo, I spent a few years in banking, then around six years in international development, before ending up as the CFO of a motorbike taxi business operating across Kenya, Uganda and Nigeria. That was my real introduction to marketplace and gig-style businesses, and I immediately saw how it is a fundamentally different proposition in developing countries than it is in developed ones. There, you’re taking informal labour — people being paid in cash with no financial footprint — and giving it structure. In our case, we had 10,000–20,000 motorbike taxi drivers who went from being completely off-grid to having bank accounts and savings products. But the dynamic in developed countries runs in completely the opposite direction — you’re generally moving people from structured employment towards gig work. In some cases this can be positive — I’ve spoken to many gig workers for whom flexibility helps them manage care commitments, deal with mental health issues or devote time to studying or other projects. But the accidental slide towards a gig economy also raises important questions about workers’ rights and protections.

When I came back to the UK, I teamed up with my co-founder, who had worked at Deliveroo from an early stage, and we started building Rodeo. The initial product was an earnings data aggregator for gig workers — specifically delivery riders, who in London might be working across Deliveroo, Uber Eats and Just Eat simultaneously. Until we came along, tracking something as basic as your hourly rate — net of expenses, across platforms — was basically impossible. The entire system is algorithmically priced, so a rider might get offered £3.84 to pick up a McDonald’s and drop it off somewhere, with no explanation of how that number was arrived at, or whether it had changed since yesterday.

We quickly acquired more than 10,000 users and tracked more than 20 million jobs equating to more than £100 million of earnings. This gave us a unique ability to understand gig worker pay and reverse engineer the algorithms. We could tell drivers things like: “Such-and-such platform tweaked their algorithm last week, and the net result is the equivalent of a 2% pay cut” — which is not information that is ever voluntarily communicated to gig workers, because they’re independent contractors, not employees. We published aggregated and anonymised data showing what riders were being paid. Unfortunately this didn’t go down so well with one of the delivery platforms. They claimed that driver pay was confidential and banned riders from using our app. They then threatened us with breach of confidence and inducement to breach of contract. We felt their case was weak but for a small tech start up fighting this through the courts was not a viable option. We have subsequently worked with the Department for Science, Innovation and Technology to make the case for a Smart Earnings Data Scheme that gives independent contractors greater control over their earnings data — an issue that is only going to become more important as dynamic, algorithmic pay becomes more widespread.

Nevertheless, that experience left me fascinated by the underlying dynamics of labour markets and the dramatic changes in how people find work and engage with businesses. This has led directly to what we’re building now.

Our current product is an AI-powered careers agent for job seekers that helps you navigate your way through the job market. You upload a CV, if you have one, and chat with an AI voice agent about your situation, your skills and your ambitions. Your agent builds up a rich picture and uses this to source everything from training schemes, project-based work or full-time jobs from a massive database of opportunities. It then helps you prepare applications, stay motivated and track your progress. We’re particularly focused on young people who are suffering the most in the current market. AI has contributed to the disruption, but we believe, when used responsibly, it can also be part of the solution.

We are still at an early stage but we have some beta testers who are providing lots of positive feedback and we are planning to do some pilots with existing career support.

Most discussions of AI and the labour market fixate on which jobs are at most risk of getting automated away, but I know you think that focus misses the more important story. Could you explain for our readers what that is?

Specific predictions about which sectors get hit hardest by waves of innovation are usually wrong. If you’d asked someone in 2005 which industry would be most disrupted by the smartphone, I doubt many people would have said taxis. History also suggests that technology displacing labour doesn’t produce permanent mass unemployment — the work changes shape, but it tends to persist. So I’m sceptical of dystopian predictions where humans no longer have a role.

What I think is both more interesting and more consequential is what happens to contracting structures — specifically, whether the boundary between employment and self-employment shifts substantially. Right now, three forces are converging that make that shift more likely than at any point in recent memory.

The first is policy-driven regulatory costs. Employment has become significantly more expensive in the UK in a short space of time thanks to the Employment Rights Act, the NICs increase and a rise in the National Minimum Wage. Crucially, none of those three apply to gig workers, but awareness about that is often sorely lacking. I have spoken to senior government officials who don’t appreciate that a zero-hours contractor — who is an employee, just without guaranteed hours — is legally distinct from a gig worker, who is a self-employed independent contractor, or that the minimum wage doesn’t apply in the gig economy. So the upshot is a large increase in the cost of employment relative to self-employment, particularly at the lower end of the income scale. A company can often get away with paying a gig worker £10–12 an hour whereas an equivalent employee would cost at least 30–40% more once one factors in the minimum wage, NICs, holiday pay and other obligations. That’s a very large incentive.

The second force is legal classification. Because there is such a big cost difference between employment and self-employment there has to be a system to prevent misclassification. Determining who counts as a worker versus an employee has always been complicated and somewhat subjective, but a series of delivery-related judgements in 2017 and 2023 have essentially cemented the principle that if a company preserves the “unfettered right of substitution” — meaning the worker can send someone else to do the job in their place — that worker cannot be classified as an employee. This has been interpreted as removing a lot of the legal risk that used to come with using gig workers and has made the model considerably easier to deploy across a wider range of sectors.

The third force is AI itself. There are two parts to this. The first is the creation of entirely new categories of work — reviewing AI outputs, training models, labelling data — that are by nature episodic, task-based and well-suited to the gig model. As AI labs improve their models they are going to need more and more specialist data that cannot be scraped from the internet.

Companies like Scale AI or Outlier are valued at tens of billions of dollars and function essentially as white-collar gig platforms, connecting AI labs with lawyers, doctors, or subject matter experts who evaluate model outputs. Similarly robotics companies are going to need massive amounts of data on how humans move: essentially turning humans into training tools. DoorDash and Uber are both experimenting with letting delivery drivers take on AI data tasks between orders.

The second part is more subtle. AI may expand the range of work that can be done by gig workers rather than employees by solving the knowledge-retention problem. Employers tend to prefer long-term employees when significant institutional knowledge is involved — when the person in finance is the only one who understands the legacy system, you need a durable relationship. But if AI can bring someone up to speed far faster, monitor their work continuously and capture institutional knowledge in ways that don’t live exclusively inside one person’s head, the case for a long-term employment relationship weakens.

One thing I thought of while preparing for this conversation was what I’d call the “legibility problem” that gig work poses — government has historically used employers as intermediaries between itself and the workforce. If that relationship breaks down, what gets lost?

Government has very little visibility over the gig economy. How we measure statistics, how we collect taxes and how we protect workers are all effectively conducted through employment contracts. The gig economy operates outside those systems. To give you a sense of the scale: there is an OECD initiative called the Digital Platform Work Reporting Directive, under which digital platforms — Airbnb, Amazon Marketplace, Deliveroo, Just Eat, and others — are required to report their sellers’ earnings to HMRC. I put in a Freedom of Information request and found that around 4 million people are earning through these platforms, for a combined total in the region of £50 billion. As far as I can tell, however, HMRC has not done much with this data, and it has been sitting with them for two years.

The implications of any meaningful shift towards self-employment are significant. There are just over 30 million payroll employees in the UK right now, and that number has been falling of late. Even a compliant self-employed person pays less tax than an equivalent employee — there are no employee NICs and more scope to deduct expenses. Compliance rates in the gig economy are, anecdotally, very low. A substantial number of people working delivery platforms simply aren’t registered for tax at all. Reducing the tax on work would be a great idea but if this happens through an unplanned slide into opaque self-employment then it is going to do serious damage to the public finances.

Pensions are a similar story. Our entire system for encouraging retirement saving is built around employer contributions deducted at source. For affluent self-employed people this probably isn’t an issue. But we are already seeing that in the gig economy — where financial literacy levels are lower and financial pressures are higher — many people will be left without any retirement plan. And then there’s sick pay, parental leave, and so on — none of that infrastructure exists in the gig economy. The employer relationship does an enormous amount of quiet, structural work that government has never really had to do itself. If that relationship weakens across a wide swathe of the workforce, those gaps become visible very quickly.

What about education — does any of this change what, or how, we should be training young people?

I don’t pretend to have all of the answers to the bigger questions of how education needs to change in the face of the rapidly evolving labour market. At the moment there is an understandable focus on AI skills although it’s quite hard to articulate what exactly that means. AI tools are also advancing so fast that the only really transferable skill is the confidence to experiment and learn.

What we can probably be confident of is that there will be a shift away from education as a large upfront investment and towards something more like ongoing, lifelong learning. Maybe philosophy degrees turn out to be highly valuable, because understanding how to reason carefully about existential questions becomes an important part of interacting with AI. But I genuinely don’t know, and I don’t think many people do, and that should be what we keep at the front of our minds.

We should also be aware not to take the role of companies in training for granted. When I left university, I went onto a graduate programme in finance. Big companies invested heavily in junior training programmes and in return people did work which was, let’s face it, often quite boring but necessary: Excel models, PowerPoint slides, meeting notes. This work has been some of the earliest to be replaced by AI and with it the need for junior employees. It should come as no surprise that graduate programmes have been some of the earliest roles to disappear.

If we’re right about the direction of travel on contracting, it may increasingly become normal — especially for younger people entering the labour market — to spend the early years of their career doing a mixture of gig work, informal skill-building, and project-based employment, before eventually securing a more stable contract. There are some industries where that has been the norm for decades. Someone going into publishing, international development, film or the arts world typically spends five years doing internships, short-term contracts and speculative projects before they’re in a position to get a ‘proper’ job. That’s not necessarily a bad thing, although I worry that it will penalise those who don’t have the luxury of family support. But it does mean that the transition from education to work looks different — it becomes more porous, more iterative — and the support structures we have for young people navigating that transition need to catch up with it.

Let’s stick on this subject and talk some more about how young people are faring. You said you think that the job market for anyone entering it right now is harder than it’s ever been. What’s driving that, and what is Rodeo trying to do about it?

It’s several things compounding simultaneously. There are simply fewer graduate jobs and entry-level openings than there once were. But beyond the headline numbers, understanding the job market has become harder than it has ever been.

Careers advice has always been a mixed bag: parents, teachers, social media, career sites give us bits of information or recommendations but they generally only see a small part of the market, and their advice is often out of date already before even considering the future. We then trawl through jobs boards which haven’t changed much since they were on the backpages of the newspapers. These systems are woefully inadequate for helping young people understand their options and find the right opportunities for them.

And then there is the application process itself, which AI has essentially broken. Applying for a job used to involve hours of tailored effort — researching the company, crafting specific answers, signalling through the work you put in that you were genuinely interested in the role. Now applicants can apply in seconds, or have an AI agent apply on their behalf passively.

The volume of applications per role has spiked dramatically. Employers, in response, are increasingly using AI screening tools to deal with the volume. So applicants now find themselves tailoring their applications to satisfy an algorithm rather than a human reviewer — which means using AI to do so. The whole thing has become a slightly absurd arms race in which neither side has much real visibility of the other.

When you look at the whole picture it’s not that surprising that thousands of young people are giving up and dropping out of the labour market altogether.

What we’re building at Rodeo is an attempt to cut through that. The model we have in mind is something like the agent that a successful actor or musician has — someone who goes out into the market on your behalf, represents your interests, finds opportunities, and brings them back to you. We think that kind of ongoing career agent, rather than the sporadic job-hunting most people do every couple of years, is going to become far more valuable. You shouldn’t be treating your engagement with the labour market as something you switch on every few years; it should be a continuous process.

Our product is available to users directly but we’re also exploring how we can work with partners to support young people who are out of work entirely — the roughly 1 million 18–24-year-olds who are currently not in employment, education or training. There’s a certain irony in the fact that AI, which has contributed to making the jobs market harder and more opaque for that group, is also potentially one of the most powerful tools available to help them navigate it.

You’ve sketched out a situation in which several significant forces are pushing in the same direction at once, and government appears largely unaware of the combined effect. What would you most want policymakers to take away from this conversation?

The central point, I think, is that the question everyone is asking — “will AI destroy jobs?” — is less important than the question almost nobody is asking — “will AI change how work is contracted?” If even a moderate fraction of employment shifts to self-employment, the consequences for tax revenues, pension saving, the welfare safety net, and the entire social contract that has been built around the employee-employer relationship are extremely significant. And that shift may well happen not because of any deliberate decision or policy, but as the accidental byproduct of three trends that nobody has bothered to join up.

Employment classification is overdue a serious review, as Matthew Taylor recommended almost a decade ago. The fact that he has now resurfaced as the head of the Fair Work Agency is interesting, but I suspect translating the recommendations he made then into legislation today will be harder than anyone would like. The legal framework governing gig work was designed for the self-employed tradesperson negotiating one-to-one with their own customers — not for someone being dispatched algorithmically by an app at scale. It is sadly another example of the gulf between how fast tech advances and how long it takes for regulation to catch up.

The more immediate thing that government could do is simply pay attention to the data it already has. Earnings data from digital platforms has been sitting with HMRC for two years. Using it — to understand tax compliance, to get a clearer picture of how many people are in this economy and what they’re earning — seems like a reasonable starting point. You can’t design good policy for a population you can’t see.

Rodeo was initially focused on reducing the asymmetry of information in algorithmic pay. The Department for Business and Trade (DBT) is currently consulting on giving the Competition and Markets Authority new powers to investigate exactly this kind of thing. Given your experience here, how do you think policy should approach it?

The data situation for gig workers is genuinely strange when you examine it closely. As an employee, you have significant rights over your earnings data — your employer can’t stop you disclosing your salary. As a gig worker, it is perfectly legal for a platform to write into your independent contractor agreement that your pay is entirely confidential. The entire body of employment law simply doesn’t apply to you.

Every gig worker is effectively running a micro-business, often with multiple clients — which is to say, multiple platforms. They need data to understand what they’re earning, which platform pays the best, what hours are worth working. They need it to file their taxes and keep their accounts. And yet they’re systematically denied it in ways that an employee with a monthly payslip never would be.

The case for what we’ve called smart earnings data — giving gig workers portability rights over their own earnings data — strikes me as quite strong, and it’s something we’ve been pushing through DBT’s Smart Data challenge process. The bigger problem is algorithmic pay itself. When every job on a gig platform is effectively a reverse auction — the platform trying to fill the task as cheaply as possible, nudging the offered price upward until someone accepts — the information asymmetry is enormous. The academic evidence on this is fairly clear: algorithmic pay has been used to extract more surplus from those transactions at workers’ expense. And the asymmetry isn’t incidental; it is the whole point. On one side of each transaction you have a company with a data science team and millions of historical data points running machine learning algorithms. On the other side you have a delivery rider who might be cycling in the rain up a hill at the time, with 30 seconds to accept or reject an offer, and a screen that roughly shows where the job is going but doesn’t tell them how long it’s expected to take.

I’m not arguing for abolishing dynamic pricing, which genuinely does help balance supply and demand. But transparency in how fees are calculated, some kind of notice requirement when they change materially, and preventing companies from actively obstructing workers from accessing their own data — these seem like proportionate responses. What I found galling about our own situation was that we weren’t trying to break anything. We were just levelling the information field. The fact that a major platform’s response was to try to legally suppress that information tells you a great deal about how much they value that asymmetry.

We ask all our guests the same closing question: what’s one interesting thing you’ve read or listened to recently that you’d like to share with our readers?

Having recently become a father, I’ve become very interested in the work of Alison Gopnik. She’s often described as a child psychologist, but in reality her work spans across human learning, the evolution of social relationship and the philosophy of mind. As AI advances, I think it will raise fascinating questions about what it means to be human — I would recommend her book The Gardener and the Carpenter to anyone interested in these issues (or who is just trying to muddle through as a parent).

To say the United Kingdom has a housing problem would be an understatement. Our inability to build the right homes in the right places is nothing less than a national crisis. The implications show up most straightforwardly in higher rental costs and artificially inflated house prices, which keep people at the foot of the housing ladder for longer and longer. Wider consequences for society include weaker economic growth and stifled rates of innovation, as labour markets become less dynamic and agglomerative forces are suffocated. Beyond these, ‘the housing theory of everything’ convincingly shows how a host of other social problems — from inequality to climate change, and obesity to even falling fertility rates — can be explained by our inadequate supply of housing.

A litany of factors conspire to create Britain’s housing crisis, but one is a chronic shortage of skilled labour needed to actually build new homes. The housebuilding industry is grappling with an ageing workforce — where around one third of workers are now aged over 50. According to the Construction Industry Training Board, the UK needs an additional 250,000 construction workers by 2028 just to keep pace with rising demand.

Britain is not alone in its failure to increase its housing supply. Numerous capital cities and other major urban areas in every corner of the globe suffer from the repercussions of low housebuilding. According to UN-Habitat, nearly 100,000 affordable homes need to be built every day to house the roughly 3 billion people who will need adequate housing by 2030.

It’s clear something needs to change.

Mollie Claypool, Co-Founder and CEO of AUAR (Automated Architecture, pronounced ‘our’), believes the solution is to change how we approach housebuilding. While the automotive and aerospace industries embraced automation decades ago, building a home remains a fragmented, unpredictable and labour-intensive process.

Rather than forcing builders to navigate high capital costs and the logistical nightmares of traditional off-site manufacturing, AUAR brings the factory to the building site. By integrating Physical AI (AI systems that operate in and interact with the physical world) with robotic micro-factories housed in standard 40-foot shipping containers, AUAR is attempting to turn construction into a predictable, software-driven service.

The goal is a huge reduction in friction. Through their ‘MasterBuilder’ software platform, AUAR automates the translation of architectural designs into robotic instructions, allowing mid-sized builders to produce high-performance timber frames on-site with a fraction of the traditional crew. If AUAR succeeds, they will not only lower the cost of housing but also provide a blueprint for how hardware and software can combine to solve real-world industrial bottlenecks.

I caught up with Mollie to discuss the mechanics of the robotic micro-factory, why the UK needs to shift from simple subsidies to sophisticated procurement, and why the most important step for any deep-tech founder working in the physical environment is to move out of the lab and face the real-world environment from day one.

What we discussed

• The transition from centralised off-site manufacturing to deployable, on-site robotic micro-factories.

• The synergy between AUAR’s MasterBuilder (the software brain) and the MicroFactory (the robotic muscle).

• How AUAR achieved a validated 30% cost reduction across 150+ feasibility projects.

• The collective action problem in the UK’s housing market and how government procurement can break the deadlock.

Lessons for scaling deep-tech construction companies

For policymakers:

• Address the collective action deadlock. Both the private and public markets are waiting for someone else to take the first risk. The government should break the deadlock by acting as the first sophisticated customer to prove the viability of distributed manufacturing at scale.

• Align grant cycles with startup velocity. Public funding in the UK often suffers from a timing mismatch. Faster tranching and more upfront disbursement are critical for hardware-intensive startups that cannot afford long lag times.

• Incentivise regenerative productivity. Link housing subsidies not just to the number of units built, but to the productivity of the build. For example, offering tax breaks to builders who achieve a specific units-per-labour-hour metric encourages the adoption of innovative approaches.

For founders:

• Prove by doing in the real world. Don’t rely on controlled lab environments for your proof of concept. Take your technology into live, messy, commercial conditions from day one. Real-world friction exposes assumptions that labs never will, and makes it easier for investors to see its viability.

• Balancing strategic opacity and transparency. While there are reasons to be wary about sharing too much information about your product, the hardware should be as legible and non-threatening as possible. Using industry-standard components builds immediate trust with insurers, regulators, and conservative customers who are wary of black box proprietary hardware.

• The advantage of multi-disciplinary teams. Bringing together people with diverse skills and expertise is a prerequisite for success. In deep tech, a team of software engineers is often blind to the physical realities of the site, while a team of traditional builders is often blind to the potential of automation.

Full interview

I. Building the Micro-Factory

Could you explain what AUAR does, who uses it, and why it matters?

AUAR builds physical, deployable, robotic MicroFactories and MasterBuilder, a physical AI software platform. Those two things together allow home builders to produce timber-frame walls, floors, and roof elements on-site without having to build or buy a factory, or buy panels from manufacturers.

The MicroFactory sits inside a 40-foot shipping container. It can be deployed in a day — that is, going from being delivered on-site to starting production — and it plugs into the 3-phase power supply that’s available on-site or provided by a generator. Our customers are charged only for what we produce on a per-square-metre basis.

What that means is that there’s no capital expenditure for builders. There’s no operational expenditure because we provide a MicroFactory production manager, and there’s no need for a very long-term volume commitment because they only pay for it when it’s producing panels. We work with mid-sized, regional home builders across the United States, the UK, Europe, and Canada who want to add automated production capacity without having to transform their business model or their operations.

Why this problem, and why now?

The pressure is coming from all directions, all at once. We have labour shortages that are chronic and getting worse. Across the UK, Europe, and the US, we need 3.7 million workers to meet housing needs. On-site crews are very hard to find and very hard to keep, particularly outside of major cities.

Additionally, the UK’s Net Zero target is adding material requirements and performance standards that manual construction struggles to meet. On top of that, housing demand shows no sign of moderating; it’s continually increasing because there’s been decades of under-supply and increased urbanisation.

We’ve been building the same way for a hundred years, but what’s changed is that the risk of not acting right now is higher than the risk of trying something new. Builders who can’t deliver faster and more cheaply are going to lose the ability to build at all because they won’t be able to keep their businesses afloat. It’s an economic sustainability problem, and governments that can’t unlock new delivery models also aren’t going to hit their housing targets. For us, the timing means we’re entering the market at the exact moment when the adoption barriers are coming down because of increased accessibility of industrialised automation.

How does that look in terms of numbers? How many people are needed on-site compared to manual construction?

It would take a timber framing crew of four to eight people, depending on the size of the house and their skill, up to eight weeks to build the core and shell of a house. With AUAR, it takes one day to produce all the elements of that house, and then a couple of days to install them with a very small crew of two to three people.

II. The Muscle and the Brain

What is uniquely hard about what AUAR is doing compared to traditional modular factories?

The hard part of what we’re solving in comparison to a traditional factory model is the deployment model itself. A traditional factory requires builders to send their designs to an off-site manufacturer, wait for production, and pay for the transport of those finished panels. That has a logistical radius where it becomes financially viable. It also means builders have to wait to be put in a queue or they have to own the factory themselves — which increases their risk exposure through very high capital expenditure and operational expenditure. Essentially, the logistics cost, the coordination overhead, and the mismatch between factory schedules and on-site work create a lot of friction.

Are there any metrics you’re particularly proud of?

We’ve validated that the cost reduction across the 150+ projects that we’ve done feasibility analysis on over the last 18 months is 30%. That’s a 30% saving on the timber framing package costs against what builders are currently paying.

Do you build your own robotics hardware, or is what you’re building primarily the software brain, MasterBuilder?

We’re building both hardware and software, but the framing matters here. We have built a proprietary MicroFactory, which is a hardware system, but we have chosen to not develop our own industrial robot from scratch. We use an ABB robot — the same kind found in automotive manufacturing — but that sits alongside a custom-designed, patented robotic system that can work non-linearly across multiple processes – cutting, assembly, calibration, routing – allowing for increased throughput, real-time control and the production of any kind of panel a design requires.

ABB has been a strategic partner for us since 2022 and an investor since 2023. Ultimately, our strategic decision was: don’t compete with the companies that have been building robots for 70 years. Focus on what’s genuinely novel, which is system integration, assembly mechanics, and software that makes the whole thing usable without an army of manufacturing engineers, and the operating model around it. You can’t separate them: the MicroFactory is the muscle and MasterBuilder is the brain.

Is there a common misconception about robotics in construction that you’d like to retire?

I’d like to retire the perception of robotics as inaccessible or that it means you have to make huge changes to how you run your business. With us, it’s the opposite. It’s a way to add production capacity when labour and cost risk is the limiting factor. Their workflow stays the same; the panels we produce are already a line item in a builder’s procurement package, and the panels are still installed by their site crews. But the manufacturing step and the data builders need for their operations is automated and predictable. They don’t need to hire robotic specialists; we deploy it, we run it, and they keep building.

III. Navigating the UK’s hardware landscape

How do you view the UK’s hardware landscape as a startup founder?

Access to public funding in the UK is one of the more difficult challenges as I see it. I think it’s also very difficult to access venture capital funding from UK-based investors. There’s a lot of sophistication in the UK around software — investors understand growth, scale, and recurring revenue. But for us, we are hardware, which is a very different conversation. Development cycles can be longer and the capital requirements are higher.

There’s a risk aversion in the UK’s startup ecosystem around funding hardware. Now, we found investors who genuinely understand that the most defensible software businesses are ones where hardware helps create the data moat. You can’t build MasterBuilder’s data without deploying and running micro-factories on real site conditions. But that insight isn’t universal across the funding landscape.

While we’ve benefited from grants — we had an Innovate UK Smart Grant that’s enabled us to develop and expand our market reach from low-rise to mid-rise housing — what we’ve found is that the timing has a bit of a mismatch for a startup. Grant disbursements can lag behind our cycles. You need either much more upfront disbursement or faster tranching.

If you were advising the Secretary of State for Housing, or Science, Innovation and Technology, what should the Government’s top priority be?

Stop waiting for the market to lead on this and use the levers government actually has. The labour shortage in construction isn’t going to resolve itself and climate targets can no longer be met by incremental improvement.

The government can use public housing programmes to encourage, for example, the use of industrialised automation and regenerative materials. They could encourage the use of distributed manufacturing. What that would mean is they would create the demand signal that gives builders the confidence to make the jump. I think the private market and the public market are stuck in a collective action problem — everyone is waiting for someone else to go first. Government should break that deadlock.

IV. The Founder’s Journey

Tell me about the journey of AUAR. How did it come about and what has the process been like?

I have worked with my Co-Founder and CTO, Gilles Retsin, for 11 years. We’ve worked at the intersection of architecture, manufacturing, robotics, and construction for our entire careers. We previously ran a research lab at UCL looking at how to develop new design methodologies for construction using automation. That lab was a founding think tank for us. From that, we won grants to build physical prototypes, which then attracted industry interest. We set up AUAR as a consultancy first to leverage that interest, learn from the industry, and bootstrap the business while we built our own product.

Gilles is an architect and technologist with a unique ability to understand design and automation from a product perspective. I come at the problem from a first principles, systems-thinking point of view. Together, we aren’t just looking to enable the industry to create good homes, but to productise the entire system.

In our team, we have a senior leadership who have scaled companies from 40 to 400 people before. Our COO, Bridget Hipwell, was Head of B2B Customer Success at Zego and has a manufacturing engineering degree from Cambridge. Our US growth lead was the COO of a regional home builder — exactly our customer profile. Our Head of Robotics was leading Rolls-Royce jet engine production for years. The important thing is we look for people who aren’t siloed in their thinking, who are agile and proactive in taking on the problems that will unlock the next stage of the business.

How has the recent explosion in AI capabilities changed how you function as an organisation?

Everyone in our team is finding operational efficiencies because of AI. For me, as a non-technical founder, it has given me the power to synthesise knowledge that exists in other members of my team much more quickly. It has also given us the ability to create very tight and efficient feedback loops between sales, product, and customer success. One person can now be 10 people in a week, and as a result, our velocity has increased exponentially.

We are a software-enabled hardware business, and we’ve always been fast in terms of hardware iteration, but we’re getting faster and more ambitious because organisationally we’re so much more operationally efficient.

What is your final piece of advice to founders building in this space?

Our attitude has always been: prove by doing, which means work in the real world from day one. Don’t do a proof of concept in a controlled lab environment. Do everything as though it is the scalable product that you will sell everywhere. Take something out into live construction environments with real customers under real commercial conditions, because those conditions just expose every assumption you have about your technology in ways that you can’t uncover otherwise. Your technology gets much more robust, much more quickly.

What is one interesting thing you’ve read or listened to recently that you want to share with our readers?

I’ve recently discovered the Centre for British Progress as I shared the stage with one of their founders recently at TxP’s Scaling Robotics event. Their research unpacks and demystifies much of the public narratives around what is really holding back the UK from progress. Their recent Substack post, “How we unleash Cræft Britannia”, is a good entry point (not least because they have announced a £60,000 prize for inventors that are pushing the boundaries of what is possible!).

Subscribe now

🌍 Philip Salter, Founder

Is Europe losing its startups? That’s the question addressed in a new paper from the EU’s Joint Research Centre. The answer, of course, is yes: around 3.3–4.3% of European VC-backed startups relocate their headquarters abroad — roughly 10 times the rate of comparable non-VC-backed firms.

The United States dominates as a destination, accounting for around three quarters of moves, with San Francisco, Boston and New York the favoured landing spots. The picture is more nuanced than the headline suggests, though, as 97% of relocations are partial, meaning firms keep operations in their home country, and in a quarter of cases the CEO doesn’t move. Perhaps most concerning, relocation is concentrated in the earliest years. Nearly half of firms leave within their first three years, before they’ve had a chance to embed locally through hiring or R&D, and this is skewed towards asset-light sectors like IT — precisely the high-value, high-growth industries Europe most wants to keep.

The most ambitious firms are the most footloose, with other research cited finding relocation rates of around 13% among larger scaleups and nearly 30% among unicorns. The primary driver is that US investors frequently require — or at least prefer — a Delaware-incorporated parent as a precondition for funding.

Whether this amounts to a serious loss is genuinely unclear. On the one hand, a Delaware flip might allow a firm to raise the round that creates hundreds of jobs back home; or it might be the first step in a gradual shift of gravity westward. Research cited in the paper suggests that around 65% of the workforce ends up in the country of relocation among firms that eventually IPO, but without US funding there might never have been an avenue to scale.

For the UK, the picture cuts both ways. Britain is the second most popular destination for relocated European startups, capturing around 7% of moves — a reflection of genuine strengths in the UK’s entrepreneurial ecosystem. That shouldn’t be surprising: our own research finds that 54% of Britain’s 100 fastest-growing companies have a foreign-born founder or co-founder, drawn from 29 countries across every continent bar Antarctica.

Policymakers in the UK — and actors in our entrepreneurial ecosystem — might be best focused on competing for a larger share of the startups that are going to move anyway. To that end, the Migration Advisory Committee is currently reviewing the Global Talent and Innovator Founder visa routes. Ease of movement won’t solve everything — there are many building blocks we need to put in place — but getting the visa regime right is a necessary condition for the UK to make the most of its position as Europe’s most attractive destination for mobile entrepreneurial talent.

Share

🚔 Eamonn Ives, Research Director

I’ve written previously for Three Big Ideas about evidence suggesting that the gig economy helps to both lower unemployment and boost entrepreneurship. This week, more data emerged that further buttresses the case for gig work. In a new paper, the authors show that the rollout of Deliveroo and Uber Eats in France between 2015 and 2019 caused a reduction in crime rates.

Overall recorded crime falls by 3% following a platform’s entry to a local labour market, but there is an especially steep decline in ‘low-skill property crime’ — such as shoplifting and street robberies. There is little impact on ‘high-skill property crime’ — such as burglary and vehicle theft — but that result, if anything, bolsters the theory the authors put forward. Gig work is disproportionately performed by young men with limited formal qualifications, or people who face labour market discrimination such as migrants. By offering these individuals an opportunity to earn an honest living, platforms reduce their ‘need’ to engage in acquisitive crime. This is standard rational choice model thinking, as first espoused by criminal economists like Gary Becker over 60 years ago, which states that when legal work becomes more accessible, the opportunity cost of offending rises.

Beyond this, the paper also shows how the spread of gig platforms correspond with a reduction in vandalism and drug crime, because, the authors explain, “[t]hese offences are disproportionately committed by adolescents and young adults and tend to be concentrated in the evening and weekend hours that delivery shifts occupy.”

As noted above, plenty of evidence now exists of the purely economic benefits of gig work, especially for marginalised people. What’s interesting about this study, however, is how it illustrates how gig work has positive, broader societal impacts too. It raises the question of what other virtuous effects such platforms might be having on society, and implores policymakers to weigh these accordingly when regulating them.

Become a Supporter or Adviser

⚛️ Mann Virdee, Head of Science and Technology

There are some technologies that seem to be perpetually on the periphery of productive commercial use. The running joke for nuclear fusion is that it’s always 30 years away. For quantum, it’s usually 10 years away. I was thinking about this yesterday as I visited Oxford Quantum Circuits (OQC), a spinout from the University of Oxford’s Department of Physics that’s about to raise a Series C.

Earlier this decade, the quantum computing industry went through a bit of a crisis of confidence and there was frequent talk of a ‘Quantum Winter’. But that’s perhaps a natural response to the quantum hype and those labelling the technology ‘bigger than fire’.

At the time, the industry was focused on increasing the number of qubits (quantum bits) in a quantum computer without the need for full error correction (which protects quantum information from errors). But it soon became clear that tackling the source of those errors, noise, was critical to opening up productive applications of quantum computing.

To be clear, there are still sceptics who believe that quantum computing cannot deliver on its promises. But a pivot towards fault-tolerant quantum computing has led to a measured and widely-shared increase in confidence about its near-term utility and the need to prepare for a world with quantum computers.

A blog last week from Google Research suggests that quantum computers could break cryptocurrencies sooner than previously predicted. They argue that there is still time for blockchains to migrate to post-quantum cryptography to ensure they are resilient to quantum attacks, but that time window is shrinking.

The deadline to prepare for quantum computers and the capabilities they will begin to unlock has been brought forward to 2029. That’s not to say we will have fully productive quantum computers by then, but that firms should be prepared.

Some applications such as fraud prevention are likely to be the lowest hanging fruit for quantum computers and can be tackled in the coming few years in the range of millions of operations. As that progresses to billions and trillions, it should open up other applications such as drug discovery.

The UK is well placed to capitalise on this. The Government recently announced funding of up to £2 billion to support the development and commercialisation of quantum technologies, and help strengthen Britain’s quantum pipeline. With the creation of our new Science and Technology Forum, we’ll be doing our part to support founders in quantum and across all areas of science and technology in growing and scaling their businesses.

Subscribe now

🤖 Philip Salter, Founder

That vast tracts of our lives are now conducted in concert with and through digital technology is taken for granted by anyone growing up today. For the next generation, the same will be true of artificial intelligence. But, for those of us living through this paradigm shift, the uncertainty can be disconcerting.

Anthropic recently published findings from a large-scale qualitative study of more than 80,000 Claude users across 159 countries. As you might expect, the headline results are mixed: 28% cite economic empowerment as a benefit, while 18% fear displacement.

I will focus on the findings most relevant to entrepreneurship and the UK, though there are broader insights worth digging into.

Globally, 8.7% of users identify AI’s primary promise as helping them build and scale businesses. This is especially pronounced in Africa, South and Central Asia, the Middle East and Latin America, where AI is viewed as a capital bypass mechanism — in other words, a way to start businesses without traditional funding, hiring or infrastructure.

Independent workers and entrepreneurs appear to be the clearest economic beneficiaries. Nearly half report tangible gains from AI, compared with just 14% of institutional employees (47% vs 14%). Those running side projects benefit most, with 58% reporting economic gains.

While the report does not drill down specifically into the UK, it does reveal that Western European sentiment is slightly less positive than the global average (65% vs 67%). Concerns in the region centre on surveillance, privacy and governance. In developed economies like the UK, “life management” resonates more strongly than entrepreneurship, with users more likely to see AI as a tool for managing already complex lives.

This last point may come to matter a lot.

Nobody truly knows how quickly AI will come to dominate, but based on recent performance, those at the most bullish end shouldn’t be discounted out of hand. This will unlock economic growth, but also disrupt much of the status quo. Those with the skills and mindset to build with AI will flourish, as will the countries in which they live.

Share

💊 Eamonn Ives, Research Director

In our latest UK AI Fieldbook interview, Murat Tunaboylu explained how artificial intelligence holds extraordinary promise for novel drug discovery.

One question I was particularly keen to get Murat’s thoughts on concerned whether or not our regulatory system was ready to handle a potential coming tidal wave of innovation. If AI does radically accelerate drug discovery, regulators could be inundated with approvals. In this world, innovation might be happening in one sense, but, in a more meaningful one, the benefits will lie dormant, waiting until a regulator gives groundbreaking drugs the green light.

As it happens, Murat was optimistic that we will avoid a backlog. But it nonetheless got me wondering about how we might be able to tweak regulatory approval systems to future-proof against sclerosis — in healthcare, and in other areas too.

One thing we discussed was the Medicines and Healthcare products Regulatory Agency’s pioneering approach to regulating the ‘processes’ involved in developing drugs, rather than individual drugs themselves. A not too far-fetched analogy here might be that if you asked a half-decent chef to cook you a meal from scratch, you’d probably trust them to make something edible simply by relying on tried and tested culinary techniques, rather than needing to forensically inspect whatever dish they eventually plate up for you.

An obvious riposte to this approach to regulation is that it could result in lower safeguards for things we consider extremely important. There may well be reasons why in some industries we very much do want to closely monitor final products and ensure they’re safe for their intended uses — be that drugs, or food, or anything else.

Perhaps we should therefore look to apply it in areas where there’s little to lose from a loosening of standards. In healthcare, this may be allowing companies working on hitherto ‘untreatable’ diseases to offer trials to patients who currently have no alternative. In logistics, it might be giving a longer leash to autonomous vehicles or drones operating in remote areas far from any human population centres. In education, it may be giving AI-powered teaching assistants to pupils for whom conventional teaching is unsuitable.

As more and more startups harness AI to advance innovation, an increasingly binding constraint on the good it could do will be our regulatory state. Now is the time to start thinking about how we can ensure it facilitates rather than frustrates modern miracles.

Become a Supporter or Adviser

🧪 Mann Virdee, Head of Science and Technology

Some scientists are deeply committed to the idea of research as a good in and of itself — that is, simply to advance our understanding of the cosmos, even when its utility in our daily lives seems limited. For others, there is a strong societal dimension — they want their research to have a strong impact, such as in tackling climate change or in improving prosperity.

But that is a false dichotomy. Research can have spillover effects in all kinds of ways that may not initially be apparent.

So, how can we measure the spillovers from science into commercialised technologies? It’s an important question because quantifying spillovers helps us to understand the social returns from science and it’s useful for those designing policy.

That’s at the heart of a new discussion paper from the Centre for Economic Performance. The challenge is that the value scientific research generates in downstream technologies is diffused through chains of follow-on research.

The authors of this discussion paper propose a new measure called Science Rank, which uses combined patent and paper citations to assign a share of the private value of patented inventions to the scientific literature they directly (or indirectly) rely on.

The authors show that Science Rank outperforms other measures, such as patent-to-paper citation counts, in identifying influential scientific research. That’s because traditional metrics only count direct links from patents to papers. This is a bit like judging a tree by looking only at the trunk. It ignores the roots, the vast network of follow-on research that eventually leads to a breakthrough.

Science Rank is more effective at identifying the technological influence of foundational research than traditional metrics. For example, under conventional citation counts, nearly half of Nobel Prize-winning papers appear to have zero impact on technology. In contrast, Science Rank recognises the value of nearly all these prestigious papers, placing them in the top half of its distribution — with over 60% reaching the top 5%

The research also shows that the US remains the undisputed powerhouse not only in generating spillovers but also in keeping the commercial value domestic. The findings provide a more mixed picture for the UK; while we punch above our weight in generating global spillover value, a huge share of that value leaks out to foreign firms — as shown by the geography of beneficiaries.

For most of pharmaceutical history, finding a new drug to treat a specific disease has resembled more a game of chance than a precise endeavour. The number of possible molecular combinations relevant to human biology runs to 10⁶⁰ — or 1,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000,000 — yet the industry’s traditional approach has essentially been to pick from that vast space at random. Compounds are synthesised and tested, and the few that show promise are taken forward, but the vast majority fall by the wayside. The process is slow, expensive and remarkably wasteful — on average, it takes around 12 years and over $2 billion in development costs before a new drug reaches patients.

Artificial intelligence (AI) offers a different approach. Rather than blindly trying to find suitable drug candidates for diseases, AI-driven drug discovery attempts to both read the underlying logic of how molecules interact and design new candidates from first principles.

Antiverse, a biotech company founded in the United Kingdom in 2017, is betting big on AI to revolutionise drug discovery and design. Its platform focuses on a class of biological targets — G protein-coupled receptors (GPCRs) and ion channels — that are associated with over 60 diseases but where fewer than five FDA-approved antibody therapies exist. I caught up with their Co-Founder, Murat Tunaboylu, to learn more about how AI is transforming drug discovery from a lottery into a design discipline, where the biggest near-term breakthroughs are likely to come, and why the UK’s regulatory posture may be more forward-thinking than it gets credit for.

What we discussed

• The key-and-lock analogy: why drug discovery is really a design problem, not a search problem.

• How AI moves from process improvement to genuine enablement — doing things that were previously impossible.

• Which diseases are likeliest to be transformed by AI in the near term, and which will take longer.

• The data scarcity problem in biologics, and how federated learning and open data mandates could help.

• Why the UK’s medical regulatory environment is more innovative than its reputation suggests — not least through its world-leading embrace of process-level approval that could redefine the economics of personalised treatments.

• The relationship between AI-native startups and big pharma — and why we might expect M&A activity to accelerate.

• How the UK is faring as a home for biotech companies — and what would make it better.

Share

Lessons for scaling AI-driven drug discovery companies

For policymakers:

• Mandate open data for publicly funded research. When government-funded clinical trials generate data — whether or not the trial succeeds — there is a strong public-interest case for requiring that data to be deposited in shared repositories. At present, too much of the scientific record stays locked in proprietary silos, duplicating effort and slowing discovery. A clear open data requirement for public funding could be transformative.

• Understand how public funding and startup timelines need to align. Antiverse credits Innovate UK funding as being critical to its survival in its early years. Matching grant mechanisms — where public funding is unlocked by private investment — can be particularly powerful, as they leverage government spending while maintaining market discipline. Expanding such instruments would help more companies survive the valley of death.

• Protect SEIS and EIS. The Seed Enterprise Investment Scheme and Enterprise Investment Scheme provide crucial tax relief to early-stage investors. Any moves to curtail or remove them risk directly reducing the flow of risk capital into exactly the kind of ambitious, long-horizon companies that the UK should be trying to cultivate.

• Address the seed-to-Series A funding gap. The UK has a well-developed ecosystem for founding companies — accelerators, angels, early grants — and an emerging one for later-stage growth. The gap in the middle, between early traction and institutional Series A investment, remains poorly served.

• Support regulatory innovation and be ready to surge capacity. The UK’s Medicines and Healthcare products Regulatory Agency (MHRA) appears to be ahead of the curve on process-level drug approvals — a development that could fundamentally change the economics of personalised medicine. This lead should be protected, signalled clearly to international investors and founders, and actively built upon. If AI leads to a wave of new drugs being put forward for regulatory consideration, regulators will need the capacity to manage that.

For founders:

• Go after the unsolved problem, not the incrementally better one. Antiverse deliberately targeted GPCRs and ion channels because they were difficult to solve. The biggest prizes in AI-driven drug discovery will go to those willing to take on biological complexity that existing tools cannot handle.

• Data generation is a core competency, not a precondition. In the biologics space, the data needed often doesn’t exist yet. Building the capability to generate high-quality, structured experimental data — and to do so at scale — is as important as the AI platform itself.

• Think carefully about international expansion. Capital intensity in biotech makes having a presence in the US almost inevitable for companies in the sector with serious ambitions. But the R&D base doesn’t have to move as well. Antiverse has kept its labs and most of its team in the UK while locating its CEO and business development activities in Boston — a model worth considering for other startups.

• Build platform before pipeline, then transition. Antiverse spent its early years proving its platform through partnerships — demonstrating that its AI could design antibodies against targets incumbent pharmaceutical companies couldn’t crack. Now it is beginning to internalise that value by building its own pipeline. The platform-first approach derisked the science; the pipeline is where enterprise value is created.

• Build capacity for grant funding. For innovative companies in need of grant funding to invest in research, do not underestimate the reporting burdens and paperwork involved. Startups may be able to bring in outside contractors to support themselves, or have staff dedicated to securing and then managing grant funding.

Full interview

I. Design versus discovery

Antiverse specialises in antibody design against challenging targets. Before we get into the science underpinning that, can you explain what it means in plain English?

This is a question we get asked a lot, and the best analogy we’ve found is that of locks and keys. Think of locks as somehow associated with a disease — where unlocking them with the right key represents the therapeutic intervention. Up until now, drug discovery has essentially been key discovery. It’s as if you are going through billions and billions of different keys and trying them one by one, maybe two at a time, with the hope that some of them will unlock the right mechanism. But for molecular combinations relevant to human biology, there are almost ten to the power of 60 possible combinations. So even if you are trying these keys randomly, you would only ever be able to sample a tiny fraction of the full key universe.

Our approach is therefore to focus more on design than discovery. We analyse the lock, and we look at other lock-and-key pairs to study how they fit together. After capturing that information, we design new keys for extremely challenging locks — the ones traditional pharmaceutical companies have been struggling with for the past twenty years or more.

How does AI make that possible? Is this essentially just about accelerating processes that already existed, or does it enable genuinely new scientific techniques to be deployed?

The first applications of AI are usually process improvement — you get better speed, quality, accuracy. Our approach was different: we said, let’s go after problems that weren’t solved before with existing technologies. And that’s where AI really enters the picture. The complexity of how these keys and locks fit together is incomprehensibly large. Our hypothesis was that if you can generate a big enough dataset of these lock-and-key pairs, you would be able to study them closely enough to infer relationships about them from within that data.

Such tools didn’t fully exist before. Creating the datasets we needed required things like next-generation sequencing, high-throughput screening, and significant automation in the wet lab. And then you also need a tool capable of making sense of that data — something that matches the complexity of the problem. That’s where AI and machine learning can come in. We can now not only generate large datasets, but we can also make sense of how things fit together in a multidimensional space. And then there’s another step: once you’ve captured those relationships, generative models can create new things based on what they’ve learned.

That’s where Antiverse found its niche. A pharmaceutical company will come to us saying: “we’ve been trying to crack this target for twenty years.” We go ahead and design molecules that bind to it. About half of what we do is in the wet lab, and the other half is in the AI.

II. Blockbusters to bespoke

Let’s say the drug discovery industry doubled or even quintupled its reliance on AI. What problems might we actually start to solve?

AI is being used across the full drug discovery pipeline — from identifying a target associated with a disease, to discovering a molecule that modulates it, through preclinical development and into the clinic. There are already many wins, and many areas are still open for significant gains to be made.

In the long run, I think there are a couple of transformative trends coming down the line. One is the shift away from blockbuster drugs — the ones that generate billions in revenue but have only modest clinical efficacy across a large population — towards smaller population groups and more bespoke treatments. AI can enable better decision-making in clinical development, so both the testing process becomes faster and the attrition rate — how many molecules successfully reach later stages — improves.

But what I think will be truly transformative is something more fundamental. In our space — antibody design against GPCRs — there are hundreds of receptors, with a few hundred associated with over 60 diseases. And yet there are perhaps five or fewer FDA-approved antibodies targeting them. It is an extraordinarily challenging area, but enabling technologies here will go far.

When you look at the later stages — clinical testing — there are multiple AI-first approaches emerging. Concepts like ‘organ-on-a-chip’ or ‘lab-on-a-chip’ allow you to test at much higher throughput at a lower cost base. Further into the future, I believe we could see approval granted not to an individual molecule, but to a process. What that means is that you wouldn’t need to go to clinical trials a hundred times for a hundred different molecules. You get the approval once, and everything that comes out of that process carries a validated safety and efficacy profile. It would allow treatment of much smaller patient groups in a way that finally makes the economics work. I believe we will eventually see truly personalised medicine — a pill not just for a disease, but for a specific individual.

Which diseases do you think are most likely to be transformed in the near term — say in the next three to five years — thanks to drug discovery techniques that only AI can enable?

I would order the likely gains based on biological complexity. Where we understand the biology better, we’ll see faster progress. So this would include things like certain cancers and metabolic disorders. Then, as you move towards psychiatric conditions, progress will be slower because we understand them much less.

This isn’t purely a problem of process or governance or technology — it is fundamentally a problem of biology. We still barely understand how all the genes fit together, how proteins interact, how cells integrate, or how disease actually emerges. We are poking around and trying to find our way, but the fundamental understanding is still at quite a primitive level. That being said, I do think AI will give us the tools necessary to probe into that biology more effectively, and that could create a virtuous cycle — faster biological understanding generating better models, and facilitating faster discovery.

III. Science in silos

You’ve spoken before about the challenge of data access in AI-driven pharma. How does that actually manifest? For instance, what data are you struggling to get, and what could you do more of if you had access to it?

AI companies are fundamentally data companies. You need good quality data that is representative of the problem you are trying to solve. Sometimes that data doesn’t exist yet — the assays or tooling to generate it haven’t been developed, or it’s prohibitively expensive to produce. In our case, where we’re working on antibodies and challenging targets, the data problem is acute, and because the field previously used a discovery approach, where you would find a few molecules and that was great, the assays were built differently and the data was collected differently.

Some of this data may exist in silos inside big pharmaceutical companies — generated over the past twenty years — but without any conscious effort to clean it, collate it, and put it into a usable database, it is unusable. Even in a quite modern pharma environment, plenty of useful data is just living in somebody’s laptop, and even here, half of it might be in an Excel sheet with the other half in a PowerPoint.

In the small molecule world it’s a bit easier — those molecules are better understood, creating data is cheaper, and automation is more mature. That’s why you see more AI companies there. When you move to biologics and antibodies, however, you are in a low-data regime: where data is much more difficult to find and expensive to produce. But things are improving. New tools are coming to maturity, and the cost of next-generation sequencing is falling on a near-logarithmic curve.

Are there any structural solutions — policy levers, industry initiatives — that you think could help?

Data is extremely valuable, and that is essentially an intellectual property issue. People are understandably not willing to share what they’ve invested heavily in producing. There have been some attempts to create shared databases where you can only benefit from the data if you contribute to it, and these could encourage more sharing.

There’s also an approach called federated learning, which I think will be quite important going forwards. Rather than sharing the data itself, you run your machine learning model on the data at the source. Whatever you take back is the improved model, not the raw data. People are more comfortable with doing that.

Another concept I’ve encountered in the United States that I think could be a big driver of change is that any government-funded research should be required to share its data publicly by contributing it to a database for others to use. Imagine clinical trials funded by the government, generating vast amounts of data — whether successful or not — and that data becoming available to the whole community. In the current situation, you often have a hot target with dozens of companies essentially duplicating the same work. Open data mandates could change that.

IV. Approver approval

Anything relating to novel healthcare technology is naturally going to be highly regulated. What is your sense of how the UK’s regulatory environment is faring? Is it keeping pace, or holding things back?

We haven’t progressed far enough in our own internal pipeline to comment on that directly. But through our partnerships, and from what I’ve been observing in general, there’s actually a remarkable story developing. More and more, we’re seeing how the UK, through the MHRA [Medicines and Healthcare products Regulatory Agency], appears to be increasingly open to a certain kind of regulatory innovation than even regulators in the US are.

In fact, there was a recent article in The Economist which focused on a company called EveryONE Medicines, based in Boston. They’re working on antisense oligonucleotides, which are molecules that can regulate protein expression. Misfolded or dysregulated proteins cause all sorts of diseases. This company was founded by the mother of a girl called Mila Makovec, who died of a serious rare disease condition. After encountering barriers in the US, she turned to the MHRA in the UK. And they are seeking something remarkable: a process-level approval. Rather than getting approval for a specific drug, they’re seeking approval for the process — so everything that comes out of it carries a validated profile. Their first clinical trial, I believe for 10 patients, is about to start. If this succeeds, it will be a first for the UK, and a first for the world.

That could be a stepping stone towards what I would call a software-as-a-drug or AI-as-a-drug track. Just imagine: you design something, and in a matter of weeks or months it gets to a human patient who has no other option — whereas the current process for getting a drug approved might typically cost $2 billion and take 12 years to complete, if it’s successful in the first place. Any mistake here would impact people’s lives in a very significant way, so regulators are rightly cautious, but this is a big change, and the UK seems to be leading it.

That’s encouraging to hear! Following on from that, another question I had anyway would be, if AI does produce a wave of promising new drug candidates, do you think the regulatory system will have the capacity to deal with them all?

The number of drugs being approved is already increasing every year. I believe it was around 52 last year. Even post-COVID, I’ve heard that the FDA has been performing well in terms of responsiveness and proactivity. And if you think about the 12-year development cycle, companies build a relationship with the FDA throughout that process — so regulators have quite a lot of advance warning and can plan capacity accordingly. If timelines shorten dramatically, then yes, the regulator could become a bottleneck, but I don’t think that’s the problem we need to solve for right now.

V. Building Antiverse

Zooming in on Antiverse itself, can you tell us more about how the company started out?

We began as one of Deep Science Ventures’ (DSV) first cohorts. It was quite unstructured at the time — more of an invitation to around 30 people to quit their jobs and see what happens. DSV has since become much more structured, to the point where it’s essentially a company factory that defines a problem and builds teams to address it.

We incorporated in April 2017 and have been building in the UK since. About two years ago, we recognised that building a really large biotech company would require very significant capital — more than we could access solely through UK channels. The US East Coast made sense for us as the place to go and establish a presence. The time difference isn’t too large, and in Boston you have Harvard and MIT, plus a high concentration of capital, and major pharma groups.

With all that being said, we made a conscious decision to keep almost everyone and our labs in the UK. As we saw it, Britain has the talent, and we benefited from good government support — so there was no need to move the R&D. But for the business development side of the equation, such as fundraising and striking partnerships, having at least one co-founder in the US felt necessary. My Co-Founder Ben Holland stayed in the UK as CTO; whereas as CEO, I took on a more mobile role and now spend a significant amount of time here in the US.

Where is Antiverse heading next? Is this a matter of scaling what you already have, or are there other avenues you think you could head down?

We are transitioning to more of a biotech company. By that I mean we will deploy a significant fraction of our capital to our own internal pipeline. We’ve identified targets associated with specific diseases, and now that we’ve validated the platform — we know it can design molecules against extremely challenging targets — it makes a lot of business sense to own those molecules and take a bit more risk for the chance of building a much larger company. This allows us to have a lot more control over our own destiny.

We will continue doing platform partnerships with a small number of select pharma companies — perhaps one over the next two years — but our main value driver will be moving our internal programmes to later stages.

How do you rate the UK as a place to build a biotech company? Have your perceptions changed at all since you started out?

When we started in 2017, I think the UK was undoubtedly the best place in Europe to build a biotech company. I still believe it’s the leader in Europe — Germany does have an ecosystem developing, France has made progress too, but there aren’t many places on the continent that can match what the UK offers.

The more honest comparison to make is with the US — where companies attract more funding and more collaboration, and have more backing. Part of that is generational — the US has been through the biotech cycle several times already, and so you have a deep pool of investors who genuinely understand the space and are looking for their next opportunity.

In the UK, we see a lot of biotech startups funded by angels or groups that made their money in real estate or fintech. They might be intelligent investors who are willing to take risk on new modalities, but perhaps without the same domain expertise to push companies forward at the right pace and in the right direction. Changing that will take generations — and it’s not something a government can fix by decree.

What policy interventions have made a real difference for Antiverse, and what else would you change if you could?

Funding from Innovate UK was, at a critical moment, extremely important for us. Without it, we might have had to shut down, or at best would have been lagging significantly behind other companies.

One thing funders should appreciate more is just how important timelines are for startups. From the moment you apply to the moment the money arrives, six months could have passed. If you miss a funding cycle, it can stretch to 12 or even 18 months. Not many of us raise enough capital to run comfortably for two years while waiting. If you are relying on grants for your company’s survival, that uncertainty can be quite unnerving. We were fortunate in that we had the ability to always treat government funding as a top-up rather than our primary lifeline, meaning we could afford to wait. But I’ve heard others describe the reporting burden as quite significant — and some companies hire dedicated staff just for grant management.

The matching grant mechanism is something I think is genuinely well-designed. You raise private capital, the government matches a portion of it. That creates a productive alignment of incentives.

A lot of Antiverse’s work involves solving problems that bigger pharmaceutical companies couldn’t crack themselves. How do you think that relationship between AI-native startups and incumbent pharma plays out over the next few years?

The large pharma companies are definitely embracing AI, and putting significant resources behind it. There’s just a huge number of things you can do with it. But their risk profile is quite different from a startup’s. Groups within big pharma tend to look for things that are achievable and make an immediate impact. Startups fill a different gap: they’re more willing to go after the moonshots, the things with maybe a 1% chance of success that require 10 years of sustained commitment. That doesn’t naturally fit a large pharma’s strategic planning horizon.

It is actually a smart play for them to watch many companies develop, wait until one becomes the category leader, and then acquire it. Maybe that acquisition costs half a billion — which might be a few percent of their annual R&D budget. So I don’t expect that dynamic to change anytime soon. What I do expect is that we will see a lot more M&A [mergers and acquisitions]. Until very recently there were almost no AI-driven biologics or antibodies in pharma pipelines. Then that jumped to around 10%, and many are now getting to clinical stages. I can totally see a future where the majority of the pipeline is touched by AI in some way — either improved or created as a novel molecule.

Is there anything else the Government could do that would move the needle for biotech startups?

The SEIS and EIS [Seed Enterprise Investment Schemes and Enterprise Investment Schemes] tax relief schemes were a significant reason why Antiverse — and probably many other companies — got funded in the first place. I know there are ongoing conversations about possibly changing or curtailing these, and, in my view, if that were to happen, it would have a very direct and negative impact on how early-stage companies get funded in this country.

More broadly, there is a persistent funding gap between the early stage, where you get companies founded on a story, and the Series A stage, where institutional investors want numbers. In that gap — between a compelling pitch and a data-driven investment case — companies often fail. Governments, accelerator programmes, and family offices currently fill some of that gap, but not nearly enough of it. That’s where most of us fail, and that’s where more attention is needed.

We ask all our guests the same closing question: what’s one interesting thing you’ve read or listened to recently that you’d like to share with our readers?

Well, I mentioned it earlier, but The Economist article about Mila Makovec is worth highlighting. Obviously the personal story was heartbreaking, and the science fascinating, but what struck me most was the regulatory discussion it encompassed. It offers a glimpse of what truly personalised medicine could look like in practice — where individual cases can drive systemic change that benefits many others. That regulatory flexibility — that willingness to adapt frameworks — is what the sector needs if we’re serious about innovation.