Tom Hulme 与 Stan Boland：从 Jensen Huang 身上学到的经验，以及如何修复英国科技生态系统

$20 trillion of value has been created over the last 50 years by building decacorns in the US. The UK has created 2, worth about $170 billion of value in the UK. So, the lack of capital crimps the ambition of companies, and therefore the best founders go to the States. We need to flood the UK with venture capital.

The biggest challenge is that for every 1 good founder, you need 5 or 10 world-class operators, and I think that's the biggest gap for us. If you look at Oxford, Cambridge, and Imperial, they're only graduating about 500 computer scientists or roboticists per year between them. We should 5x that number. If you graduate in engineering or computer science or something here, you should have a Tier 2 visa stapled to your graduation certificate. Ready to go.

Guys, I am so excited to make this happen. Two of the smartest people, I think, in European and UK venture and startups. I want to start with a little bit of context. Stan, if we start with you and then move to Tom, I love listening to you speak about the UK and where we are. What's the background as to how you got here, and just the quick 1-minute intro on you?

Yeah, so I joined a company called Acorn, a computer company based in Cambridge, back in 1997. It owned 40% of this company called ARM, so I helped get ARM public and then figured out what to do with Acorn. I set up a chip company out of Acorn, which got venture funding. We raised $30 million of capital and sold that company to Broadcom for about $640 million about a year and a half later. It was an amazing deal.

I did a second deal in the chip space, where I built a company and sold it to NVIDIA, and did a third deal in the AI space. I've serially founded, run, and then sold companies, raised about $330 million in venture capital, and sold them for about $1.3 billion. So, that's what I've been doing for the last 25 years.

Well, I'm here just to learn from Stan. That's the reason I'm here. Tom, what about you?

So look, I won't give a long bio, just a really quick one. I helped set up GV in Europe the year you started 20VC, in 2014. We've now done over 50 companies and invested in 12 countries. We just broke through half a billion dollars in the UK alone with our investment in Isomorphic Labs last week, and I'm passionate about making the European ecosystem as vibrant as possible. I'm keen to discuss that.

I want to discuss it in a way where we're going to cite the problems and then cite the solutions. I don't want to be Debbie Downer and just do the problems, but I also want to be pretty granular on the solutions.

I think for me, the biggest problem is actually talent supply and not being a magnet for the best developers in the world. London or the UK isn't anymore where I think it maybe once was. Do you agree that we have a fundamental talent problem today in the UK?

I think we've got a bit of a talent problem in the UK. I don't think we're the magnet that we were or that we could be. It's quite interesting, actually, if you look at where talent is being born in AI across Europe and where it lands in terms of where it stays. The UK is minting about the same talent it's keeping, but that is net-net, actually.

We're losing talent to the US, and we're recovering some of that from other parts of Europe. Net-net, we're about the same, but we could be 10x better, frankly. That's the key point: we ought to be making the UK the magnet, the place to set up a company in Europe. All that talent that's leaving the UK and other parts of Europe to go to the States—we ought to be capturing it and building companies here.

I'd say we're losing a bit of it.

Yeah, I think of it in terms of engineering talent, and then I think of it in terms of founding talent. How do you think that differs? As you said, net-net for deep AI engineers, my worry is whether we actually have an exceptional founder supply that maybe other countries do. That's the difference I think about from my perspective.

I completely agree. I think we're rate-limited. It's the biggest rate limiter, actually: the supply of founders and the supply of operators. The great thing about founders is that they'll smash through walls to build stuff. You have Melanie at Canva, who built that business in Perth, Australia—no right to build a $50 billion business in Perth, but it can be done.

If you gave me the choice to have more Nicholas Zennströms, Demis Hassabises, or Stans, I would absolutely take that. I think it could only be a good thing. The biggest challenge is that for every 1 good founder, you need 5 or 10 world-class operators, and I think that's the biggest gap for us. That's the rate limiter.

To Stan's point, if I just look at engineering talent, we've got 3 of the best 10 universities on the planet here. If you look at Oxford, Cambridge, and Imperial, they're only graduating about 500 computer scientists or roboticists per year between them. We should 5x that number. There's a huge demand. I don't see why we aren't increasing it.

To Stan's point, we can do a better job of making it appealing for the most entrepreneurial talent to come into the UK and maybe retain the talent that does study here and becomes expert.

The 2 big exporters of talent in the world, I think, are China and India. The majority of the graduates there have decided to go and work in the States, frankly. Even if they come to university here, they're typically not staying. They come here with pretty much no intent of staying. In fact, we're not really welcoming them either.

If you graduate in engineering or computer science or something here, you should have stapled to your graduation certificate a Tier 2 visa and rights to stay, as well as the right to bring your family across. Just make the UK the place that people want to come to. That's what we should be doing.

I love that. Can I build on that? I think you become what you measure, and the government is measuring a lot of lagging indicators. We invested in Stripe in 2017, and one of the things that struck me is that the Collison brothers were tracking a KPI. They were tracking the number of Series A companies that transact online and that were using Stripe, and the number was phenomenal. It was in the high 80s percent.

Taking Stan's idea, our government should actually be looking at the people who are graduating and asking what percentage are choosing to stay. That is the leading indicator. Great founders focus on leading indicators, not lagging ones.

Totally agree with you. You mentioned attaching the Tier 2 visa to the graduation ceremony ticket. Is there anything else that we could do to make sure we have a high talent-retention number for great engineering and founding talent?

I think the second big factor is money, which I'm sure we're going to talk about in a second. Money is—there's no structure to this. Yeah. I think money is the great attractor of talent as well. Part of the reason that people will come to the UK, come to London, or come to the Golden Triangle is the fact that they can get funded here.

They can not just get funded at pre-seed and seed, but at Series A, Series B, and Series C, through the growth phase as well, and in fact keep the company here. The constraints that come from a lack of capital are also a factor.

The model in the UK has really been, "Let's build early-stage companies, get them to a certain point, and then flip them to America." A lot of founders might be thinking, "Why don't I just skip that first stage? Why don't I just jump on a plane and form the company in the US?"

Why do you think we have a lack of capital in the UK?

I disagree with you, so I'm intrigued why you think we have a lack of capital.

You just need to look at the numbers. The numbers say that the model to copy is the US. The US is so obviously successful in technology. $20 trillion of value has been created over the last 50 years by building decacorns in the US. The UK has created 2, worth about $170 billion of value in the UK. So, we're 2 orders of magnitude off the US.

The US is a model to copy, and if you look at how much venture capital was raised by US VCs last year, it was about $76 billion. Pro rata to population, the UK should be at $15.4 billion. UK funds raised $3.7 billion last year, so we're short about $12 billion in venture capital.

I absolutely hear you. But as a day-to-day venture investor on the ground trying to find companies and great people to invest in, there is simply not the supply of entrepreneurs if I were to keep my bar as high as it needs to be to build great companies and deploy that money.

I think there's a chicken-and-egg situation here. Traditionally, the way to think about this is that you create momentum by building successful companies. The idea is that capital flows to places where it gets a return. Therefore, you create a track record of building companies here, and capital will flow to the UK. That's the causality.

I think the causality is actually the other way around. If we put capital in place here, great companies will rise to the occasion, and the supply of companies will come. The reason I say that is that there's a country you can look at where this is true, and that country is China.

Twenty years ago, China had pretty much nothing in technology. The Chinese studied the US model and put huge amounts of capital in place, and now China is clearly global number 2 in terms of technology.

You look at the amount that's invested in AI, for instance. There are only 2 countries really investing in AI: the US and China, and the European investments are diddly-squat. You almost can't see them; they're that small. The net result is that we've got a very successful Chinese tech sector.

I think the goal should be that the best capital gets concentrated in the best companies. China is an amazing example: you get concentrations of talent and then concentrations of funding taken to an extreme there.

I think one of the data points that makes this so difficult is that none of us would think all companies should get funding. The real challenge is that, if you ask any founder—and by definition, at seed stage, maybe the majority shouldn't get funding—when they don't receive the funding, they think it's a funding gap. I don't think what we should be doing is necessarily just evenly distributing capital across the whole market. I actually think that's damaging for talent concentration as well.

Instead, we should have sophisticated people who say, "These are the companies that can win. These are the companies that can actually absorb more capital because the founders are great. They're not going to be overcapitalized. They'll then bring in the best people, and maybe they can just be more ambitious."

I've got a good example of this, actually. There's a company I've invested in called Wordware—a good name-check for them. There are 2 guys who studied computer science in Cambridge. They could have set up a company here, raised probably $5 million on a $20 million pre-money valuation, and built a set of tools for LLM prompt engineering. But they went to San Francisco instead. This is Philip Kazer.

They ended up raising $30 million on a $220 million post-money valuation. Those investors are expecting them to build a business worth $2 to $3 billion. That's 10x. So, even better. I think those investors are expecting him to build a $10 billion business. At 10% ownership, you need a billion dollars. But the fact is, they've got the capital to do it really well.

That stratospheric raising of expectations is part of the US playbook. The provision of capital behind founders with energy and enthusiasm does work. I completely agree, Tom, that concentration really matters: the ability to put a large amount of money at the right point behind founders who have the energy, intellect, pivotability, and coachability is absolutely critical. It's the part that's missing, I think, in the UK and in Europe as a whole.

I don't think we need more money. I'm seeing every day the most inflated prices, and it's because you see this concentration of capital to obviously good people, like your Wordware, where you can get a $5 million round at a $30 million valuation, and then Lightspeed and General Catalyst come in and suddenly it's $6 million on an $80 million valuation. It just goes nuts.

I see the complete removal of liquidation preferences now, and it's because we don't have the supply. The capital concentrates and just inflates in a way that's much more significant than in the US. I think we have this fundamental talent problem, and then we have a narrative problem based around the behavior of venture investors in Europe.

If you speak to Filip, he'll tell you that it was super fast in the US. They totally got him and gave him a great experience. In Europe, it takes weeks, the partners aren't there, and they're slower. We have a very bad customer experience for founders in Europe, which I think makes it a less attractive funding product than the US.

I think that's certainly true. My solution would be: let's increase the amount of capital here, and the best founders will seek out the best VCs. The best VCs will generate outsized returns, and they'll be able to raise the next round of capital. You will gradually, and hopefully quickly, ratchet up the performance of venture in Europe.

Actually, the thing that has scared me historically, when people have talked about, for example, government investing in startups, is that I think it's an incredibly difficult thing to do. I don't know if I'm any good at it still, because the feedback loop is probably a decade. It's the worst learning loop ever.

The important thing is to make sure that, if there is more capital in the system, it's deployed by the experts.

Correct. It would be an absolute disaster for the government to be making direct investments in companies. There's no way they can do it.

If we want to get really spicy, though—and Tom's seen my Twitter, and I don't give many shits anymore—most of the British Business Bank's portfolio is just dire. These funds should not be in existence. The question is: do you have the right to win? Do you have the right to find companies, pick them, win them, and help them get better? The majority are honestly dire, and they will not do well. Government money will be wasted.

If you want this to go to truly gifted individuals who will invest it wisely, then we should see real concentration of capital into 3 to 5 players in the UK, because honestly, I think that's the number that is really good.

A couple of quick reactions. Firstly, I don't think it needs to just be 3 to 5 players in the UK. It can be global funds. I think you have some of the best. Secondly, the best funds have proven themselves over multiple vintages now. They're oversubscribed, but I would hope that UK plc could get into those funds.

I would say that, firstly, no large fund of funds has ever lost money. From an investment perspective, I think the government ought to be willing to take a much bigger risk on fund investments here in the UK.

I think the British Business Bank puts something like $424 million a year into fund investments, which is a drop in the ocean compared to the $15.4 billion that we ought to be investing. That number needs to be 10x in my view.

Secondly, I think there is a venture talent pool that can be energized. Below partner level in a lot of these firms, there are a bunch of people at principal level who could be interested and willing to run a new fund, and would do a bloody good job at it.

I also think we're at a time when US partners would consider coming to Europe if the capital was available. There is talent in Europe, and valuations are lower. If you could put the money in place, I think we could not only release some homegrown talent from venture firms, but also imagine some of the leading partners in US firms coming to London or the UK to get this economy really moving.

Sometimes in my head I think, "How many friends do I want to lose in one single show?" My question to you is that I don't agree that prices are better here, honestly. For the best companies, like your Wordware, if they were to stay, they're just super high. They're so inflated. I think—just a quick thought on—

No, no. If I look at where we sit today, some of the best deals are overpriced. I think it's often because they're the ones with the traction, and they're therefore somewhat de-risked.

There are 2 things that make this a really difficult thing to answer. We talked about lagging indicators. The first is that we're working against sources of capital that were raised in the past. These are often not brand-new funds, and they were often raised in a zero-interest-rate environment. The cost of capital has gone through the roof, given the current interest-rate environment, and I think that's going to get worse, if anything.

The fact that a lot of these funds are giving out so many stock grants means you basically need to hit a 20% IRR to break even. These numbers are really high. I actually think there's probably going to be less money in the market for venture in 2 years than there is today.

The second thing is that, with classic machine learning, I think we're overfitting to history. I don't think we know what the biggest companies will look like going forward, so it's very difficult for me to say that the returns profile funds got from investments 10 years ago will be the same going forward.

My belief is that AI is creating a real power law, far more than we've ever seen before. The job to be done is going to be being in those handful of global champions.

If you look at Israel, I think it's an interesting example for us at the moment. It's an amazing story: recently, the Wiz acquisition was $32 billion. That's about 7% of Israel's GDP. A lot of that is actually flowing back to Israel, and it will create this multiplier effect.

That business was basically built in 5 years. It was assembled without actually having a clear problem identified. They just got a world-class team and capitalized the business really well on day 1.

I think the businesses we want to build look more like Wiz, and so we should concentrate capital into the best founders. Can that be done from the UK or Europe? Hell yes. What we do often at the moment is say, “Be close to your customer.” We say, “Go to the US because the market size is roughly an order of magnitude bigger than it is in the UK.” We’re not saying give up the US market. Absolutely go to the market, but build a global business on day 1.

Yeah, I think that’s right. I think it’s almost pointless building a number 3 or number 4 in the marketplace today. If we’re going to undercapitalize businesses and build businesses that are number 3 and number 4, it’s not what we need, because those businesses have got no choice but to be sold to US companies.

We’re never going to create companies here that stand up on their own 2 feet and generate the jobs growth and the diffusion of wealth that the country desperately needs. So I think we’ve got to concentrate on companies that could be global number 1 or global number 2, which does require big checks to be written to those companies at the right point.

Can I give another example of this? In a way, we’ve got a problem that we’re subscale in the way we’ve described it. I agree with that. The other place where I think our relative size hurts us is in subscale pension funds, for example. You’ve got 90 local pension funds.

Actually, a policy that I was really excited about, which I think the chancellor mentioned last year, is this idea that they should be aggregated so that they can have a world-class investment office. They can do something like Yale, like when I do LP calls for emerging managers.

Thanks, dude.

You’re very welcome.

Tom had to do like 10.

I think it was more.

Literally, I just told you about the 10.

No, I did do a few. The thing that’s stunning about the US firms, and then the really sophisticated ones here, like Wellcome Trust—just phenomenal investors—is that they understand the power law. They understand that they’ve got to build relationships for the long term, and they can actually have world-class analysts inside those firms. You can’t expect a tiny fund to do that.

So this idea that we might aggregate 90 local pension funds in the UK to enable them to think more like Yale, rather than just replicating the asset split, I’d be really excited about.

I thought it was so interesting that you said it doesn’t make sense to build these number 3 or number 4-tier players in a market, because I’ve been in venture for 10 years now. A lot of the job has been, “It’s like HR platform X, but in Europe it’s Y.” Actually, you can build billion-dollar, $2 billion, or $3 billion companies on the back of that.

Where can the UK and Europe then be a number-one market leader and beat the US and China?

Well, I think if you think of it as a stack, from semiconductors and hardware up to the applications layer, then I think it’s easier for Europe to think about building at the bottom of the stack or at the top of the stack. I think it’s quite hard for Europe to build in the middle of the stack.

I think AI application companies that are solving a particular problem, particularly if there’s a defensive moat that exists in Europe, are obviously a good place to start. At the bottom of the stack, I think something that’s close to the metal—semiconductors that are solving a particular problem—happens to be somewhere where we have the expertise to do that. It happens to be a B2B sale where we get paid for the value of the architecture that we put down and the utility it delivers.

I think it’s easier to think of the top and bottom of the stack as the places where we can build those companies. It’s not necessarily where we’re focused, but it is where we should be focused. Whereas if you’re building some middleware layer or some tools layer, I think it’s a little bit easier to imagine doing that in the States than doing that here.

I think the interesting thing with Aatish’s argument is that I really agree with it. I like the idea of focus and specialization. One of the things that concerns me is this idea that we can be experts at everything.

Instead, I think we have to understand our unfair advantages. If, for example, the bottom of the stack, the infrastructure layer, is somewhere we can be world-class, we’ve certainly got the technical talent. Then I think we have to build the whole ecosystem and structure it and say, “Actually, in this one location, we’re going to be effective.”

We then have to do second-order things. We have probably the highest electricity or energy costs in the whole of the Western world in the UK, which just does not enable you to do a great job of this. It doesn’t even enable you to do a great job of training foundation models. If the blended cost of training a large language model is 20% energy, we’re already kind of losing.

The important thing is to say, “Actually, what are we going to be world-class at, and where are we going to be?” We have some advantages. One of the things that’s interesting—we’ve done it in this conversation—is that it’s easier to aggregate everything at the national or continental level. In truth, we should be honest that London is incredibly different from the rest of the UK.

Building a startup in Europe is doing it on ultra-hard mode. We’ve talked about it before, but if you do it in London, it’s slightly easier mode at the moment because of the talent and because it’s where the investors are. So we have to start to acknowledge that, lean into it, and actually have these pockets of specialization.

Yeah, I think that’s right. I wasn’t so much thinking, by the way, of building lots and lots of data centers on expensive energy costs, because that would be nuts right now, obviously. I was more thinking about the chip design layer.

So, not even chip fabrication, but chip design, which is where 75% of the value in the semiconductor space is. NVIDIA, Qualcomm, and Broadcom are all basically semiconductor design companies that sell chips, but get them fabbed by TSMC or whatever. That’s the model that we ought to be playing in. We have something like 2% of that global market in Europe. It’s insane, honestly.

In the fabless space, we must be building successful fabless companies, I think. Europe—and, in fact, the UK, in Bristol, as it turns out—happens to have this full-custom microprocessor design capability that stems from the creation of Inmos 40 or 50 years ago, which is kind of unique, actually. There are probably only 2 places in Europe where you can do that, and Bristol happens to be 1 of them.

I think it’s plausible to build companies in this space that are global winners. You’re right that we do need to put much larger checks into those companies, but that’s the reason why we need more venture money here: to be able to write those checks.

It’s interesting that you said that about the cost of energy. I was speaking to the CEO of one of the largest data center providers in the world, and he said, “Harry, in the US, my energy costs are 4%. In the UK, if I set up today, it’s going to be 17% in total.” I was pushing and pushing, and he said that. I thought, “All right, fine. You do you.”

My question to you then is: when we look at that and we look at the money that’s needed to fund it, where does that money come from? I understand your argument around the scale and the scale of cash needing to change. How do we fund the 450 million that BBB does invest to whatever we want to call it—$2 billion, $3 billion, $4 billion?

Well, firstly, I think Europe has a lot of money, actually. So I guess the first thing to say is that Europe’s got a lot of money, obviously in pensions. We talk a lot about pensions. It’s got a lot of money in family offices that are locked up all over the place, actually. So Europe is not capital-short; it’s just not investing in this particular asset class.

The job, I think, of BBB is to create that asset class at speed and to play an enabling role in doing that, essentially. My suggestion would be that we get the government to increase the amount that British Business Bank puts in. We may need to uprate the quality and talent in BBB to be able to do this, but BBB puts like 4 billion a year in and would require a 50/50 funding ratio. The GPs have to raise matching money; otherwise, BBB doesn’t participate, but it can be 50%.

If I want to create a billion-dollar fund, I know I’m going to get half a billion from BBB, and I’ve got to raise the other half a billion. Raising the funding ratio to 50/50 would be a good start.

Then I think we’ve got to be creative, which I guess is another call to action for BBB, about how we split the fees and split the carry between the different LPs in the fund. At the moment, there’s a lot of hand-wringing and anguish about the fact that pension funds won’t pay a 2% fee. I would say, “Fine, let’s do it on a 0.5% fee then.”

Instead, the carry that the partners have is higher, and quid pro quo, BBB might pay a 3% fee and the carry for the partners is lower. But net-net, we’re still at 2 plus 20. So let’s be creative about how we do it.

The job is to bring the capital in and make it mesh with public money to mint these large funds that can write these big checks, allowing us to play seriously in some of these sectors that are basically capital-intensive and winner-takes-all. That’s what we need to do, I think, to pull ourselves out of the nosedive that the country is currently in.

I think, where would we get that money from? Oh, well, the government has created its own fiscal freedom to do this, actually. The government is able to treat any investment in BBB money as not borrowing or public spending. It forms part of public-sector net worth, and it doesn't count as current spending because the argument is—and I think this is correct—that what we're doing is building up a financial asset on the government's balance sheet.

If you did this consistently over 10 years, you'd have 40 billion of fund-of-funds investments in venture on the government's balance sheet. The worst-performing funds of funds generate maybe 6% IRR; the best-performing generate mid-20s, so it's always higher than gilt yields.

I would say you could go even further. You could say, “In 10 years' time, we've got 40 billion on the public balance sheet. Why don't we make an offer to the public?” Why don't we offer it to individual pension plans to invest in this stock? So, yeah, we could create a fetcher moment, really, where you privatize.

People in their 20s and 30s should be owning assets in the future of the country, actually. They should be owning those assets, and they should be recycled into making the country more successful competitively. Technology is the place to put it, obviously. That's kind of what we ought to be doing.

I do see it as investing. We're talking about infrastructure projects. We look at Germany's trillion dollars, and I think it's incredibly important. I like the idea that we have a kind of intellectual infrastructure investment that you're describing.

The big thing to design around—and it sounds like you've started to think that through—is the adverse-selection bias. My biggest fear, because there's such a power law of returns, is that you don't want to just end up with the worst investors making the worst investments. Placing an emphasis on those—maybe first supporting first-time funds and solo GPs initially to get going—could make sense, but it's incredibly important for the UK taxpayer to get into the best funds.

I do believe there have to be incentives that UK plc can provide so that the best funds that Harry describes are actually excited to take money from that BBB fund of funds. But Tom, do you think if we put such a system in place and made it plausible and feasible for GPs to go and raise a half-billion-dollar or $1 billion fund here, we'd get partners in US firms with a strong track record to consider coming to London to basically raise a fund here because it can be done here?

So, I think the answer is yes. But again, it sort of speaks to specialization, and I guess the question for me would be: in what areas would you get the best people saying, “It's worth me doing that”? It wouldn't necessarily be in digital health, where the UK has one major customer and no one else.

It would be in places like fintech, where we have a good track record because we're in a great position globally at this point. That's why we've done a disproportionate number of fintech investments. Defense, I think, is an interesting area at the moment, where we're going to have to look more to 3% of GDP spent on defense. So there'll be areas where I think very smart, rational people would make that call. But there are others where it would be a harder stretch, like consumer, where it doesn't really make sense to be outside one of the biggest markets.

Why do you think defense is different from health?

I think in defense you still have 1 primary buyer here, really, which is obviously the MoD. Then you have very splintered and fractured buyers, which is the rest of Europe, and each wants to have its own dominant domestic provider.

Disclaimer: I'm a reservist, as you know, so this is something I'm really passionate about. I'd say there are 3 things happening at the moment that make it significantly more interesting than it has been in the past. The first is that very smart people are interested in doing it because they think it's right. There are people like our peers who are interested in starting defense companies because, for the first time, they actually think there's an existential threat.

The second thing is that, while you do say—you're right—there's maybe a single buyer, it's more complicated than that. In the UK, we have multiple services and multiple regiments within each; each is a potential customer, and they're being forced to innovate at the moment.

The final reason is that, to some extent, we are geopolitically close to a war zone at the moment, and we have a point of view in that war. We occasionally have some of our armed-service personnel at risk. I think those 3 things together mean that, when you look at Anduril in the US and its recent round at $8 billion, which was oversubscribed, it shows you there's an appetite among people and capital to go in there.

I think the UK has interesting talent. The UK is playing its part in Ukraine at the moment. It's an amazing place to test new technologies, and I think it's an opportunity to build next-generation primes here. So, as a category, I think defense in Europe is an important one at the moment, and there's probably $2 trillion to $3 trillion going to be spent over the next 5 to 8 years in Europe on defense, actually—all layers, not just the final product, but components as well. There are lots of layers here.

I agree, and I do think that will forge some dual-use technologies. If you look out there at the biggest defense companies, you could argue that DJI is one of them at the moment. Actually, I think you'll see the same thing in reverse. Some of the technologies, whether it be cyber or maybe UAVs—drones—I think you'll start to see them have other applications outside the military.

When we think about the kind of amazing companies you've mentioned—and we've mentioned some other amazing ones—in the US, there is a market for them to go public. There is a liquidity market that is much more vibrant. In the UK, we have the London Stock Exchange, where a lot of people throw a lot of criticism, and people choose not to list on the London Stock Exchange. To what extent do we need local domestic liquidity markets, or are we in a global world where you can just go to Nasdaq?

I've thought about this a bit, actually. I think it's a supply problem again. The lack of tech companies in London is stark: there's only 1 London-listed tech company worth more than 10 billion, and that is Sage. Sage is a 30-year-old ERP company. It's a very nice company, but as an output of the 20 billion a year that we pump into tech in the UK, to have 1 company worth 10 billion on the stock exchange is not a great outturn, really.

Whilst the US has minted 20.5 trillion of value in its tech companies, we've minted about 100 billion over that period of time in tech. So, firstly, let's accept it's not good. But I think the problem is supply, actually: companies grow to a certain size and they're stunted for all sorts of reasons. It could be quite early on—the cap table's broken, they hire the wrong people, or they have the wrong product-market focus—but it could also be a lack of swinging for the fences, a lack of money to swing over the fences, actually.

The net result is that companies just have to be sold, typically to US buyers. So they never get to the point where they grow and are capable of going public. There's not a big pipeline of companies coming through that could IPO. There's a handful in fintech, maybe, but apart from that, not very many.

So, I think it's a supply problem, actually, and that's why it's really important that we grow the amount of capital here—UK capital that is patient and will put the money in—so we can fund the companies all the way through to eventually going public. Then I think it will be natural to list them.

Yeah, where there's a market for them. I think that could be London, it could be Nasdaq, it could be wherever is suitable for the company.

Agreed. Definitely, the supply problem doesn't help. If we had many more, much bigger companies, we wouldn't see it. I'd give 2 other reasons.

The first is a sentiment problem. I have not spoken to anyone for months who is positive about the LSE or listing, whether it be valuation or perceptions about, for example, the product itself because of the stamp duty driving down liquidity. I'm afraid these stories are kind of like SEO for our minds. We hear the story, we remember it, and there's just a negative sentiment about it.

Most good companies are becoming more open to the US, and they're getting courted very effectively. They have the red carpet rolled out for them. So that's the first one: the sentiment problem needs to be turned around. I mean, you interviewed Julia Hoggett. I don't know your point of view, but the sentiment isn't great.

The other one I'd just point out is that I think it's an easy thing to measure. That doesn't mean it's the best thing to measure, actually. If I'm completely honest, given the choice between picking where a company's HQ is, where the bulk of the employees are, where the IP is generated, or where it's listed, I'm taking the first 3.

They're way more valuable. I know that they're kind of interlinked, but the most important thing is where the economic driver is, where the employees are, and where that value is created. If we do have a period where the very best UK and European companies end up listing in the US, I think that's okay, as long as we have a great platform of big value generation here. I think it'd be okay if the ownership of those companies, when they go public, is predominantly here in the UK, because I really think we've got to set a national goal here for wealth creation.

The UK, really, it's clear—you just look around—and the country is getting poorer. We can't afford all the services that we want.

So what do you mean by a national goal for wealth?

Firstly, I think tech and innovation is really the engine of economic growth here. There's no other engine that we can rely on. If you look at the US, it's created 20 trillion of value over the last 20–30 years in new tech companies. The UK has created 0.1 trillion, so, pro rata, we should have created about 4 trillion. We've created 0.1 trillion, so we're about 4 trillion short of where we should be.

I think we could set a goal and say, “What if, in 20 years, we set a national goal of creating 4 trillion of wealth in tech?” That's obviously a sort of escalating growth in value. Let's say, at year 10, the goal is 0.5 trillion, and thereafter we grow from that point. Growing 0.5 trillion is already quite a big goal for us, given that we've only created 100 billion right now.

It also sets the mindset for saying, “What are we going to have to invest to do that? What do these companies look like? How much capital are they going to need?” They're going to need about 100 billion of capital to do that, realistically. You think, “Okay, that 100 billion—where's it going to come from?” It's going to be something like 10 billion a year that we've got to put in, in addition to what we're currently doing. That's roughly the gap in our venture capital.

SEIS and EIS have been very effective in encouraging more direct investing from individuals.

When I look at my cap table today—or my LP list today—85% of dollars, maybe 90% of dollars, are from the US for me. I'm thrilled and honored to have them, but it is slightly—not alarming, but I think about it—that we'll do very well, our funds will make a lot of money, and all of that will go straight to the US.

That doesn't thrill me for my grandparents, who have pensions, and my mother's got pensions, and everything around us in the UK. Is there anything that could be done to unlock the huge amount of family office and corporate pension fund money to invest directly into funds, whether it's an SEIS for funds or an EIS for funds? Otherwise, they're not freaking moving.

I think the BBB role I spoke about earlier is critical to this. If you look at where the money came from in the US, and the distribution of where that money came from, it's pretty evenly spread across endowments, family offices, pension funds, insurance companies, and so on. It's not just pension funds. There are other sources of capital that we need to energize and create.

We don't have the endowment fund pool, that's true, but we do have more family offices. I think there's a lot of money here. There are 1,100 family offices in London.

Blimey.

Yeah, it's a lot. I met every one of them, which is why I think we need an energized BBB that is creative about the structuring of deals to bring those people in and structure them in a way that makes it easy for them to participate in this illiquid, 15-year asset class. The fee structure and the carry structure need to work for them and work for BBB.

You'd end up with LPs that are 50% the national balance sheet and 50% UK-based pension funds, endowments, family offices, and insurance companies.

Listen, I'm spending more and more time with politicians now, and they're all just terrified of getting fired and terrified of headline risk. When I listen to you, I'm like, “Great. I see all of this,” but then I see the Daily Mail headline, which is about how your taxpayer dollars are going to fund Tom or Sarah's venture fund, where they have a Porsche and a nice house in Hampstead, and the concentration of wealth on your taxpayer dollars.

Do you think we're actually being reasonable in thinking we can do that? Do you share my concern about that headline risk?

It is definitely a challenge. I see the challenge. But I actually think we've got to make the case for why the UK needs to change. We're not really fulfilling our potential right now. We've got a lot more to achieve.

It's about raising everybody's sights to build this country to be the best it can be. Let's build this value that's missing in tech, because it's not coordinated right now. This is 20–30 billion a year that we pump in at the front end, per annum, in tech—150 billion over a parliament in university funding for science and tech, in SEIS, EIS, VCTs, R&D tax credits, patent box, and so on. You add all those things up, and what's coming out of the pipeline is—nothing, really.

Some people are making some wealth along the way, but that's not what we want. We're not achieving a national goal, really. If we say, “Let's do this together as a country. Let's build this value and let's energize people,” it's clear to me that active money is the way to go. Passive money is not the way to go.

Active money means that when things are going well, investors double down. When things are not going well, they kill it. We've got to be courageous enough to do that. That does require—I mean, VCs require OPEX cover, don't they? So you've got to basically fund them, really.

I think 2 ideas that Aatish's thoughts remind me of. The first is, one of the things I admire about Sequoia is that its meeting rooms are named after its LPs. I think that's a really interesting thing to remind everyone who they're in service of.

One of the challenges we have in the UK is that we perhaps don't celebrate entrepreneurs as much as we might. If we were able to say to those entrepreneurs that they can tell the story about the wealth they've given back, whether it be through BBB or another vehicle, I actually think the public would see more of the value they're creating.

The second story I think about is the Norwegian sovereign wealth fund. Extraordinary business. If you look at their ownership at the moment, it's mind-blowing. But the other thing they do is they effectively have a stock ticker, so that everyone can see in real time what that national wealth is. They have a literal stock.

I interviewed him, and he's literally like, “The happiness of the country does go up and down dependent on the ticker.”

Exactly.

So this is all about just reminding society that some of these great entrepreneurs are building businesses in society's service. I think that's what we've not done.

Great idea, actually. If we have this 4 trillion goal, it'd be a great idea to have a national ticker as we climb our way towards it, wouldn't it?

I think it would glue culture and society a bit more than perhaps you have at the moment, where it's perceived to be haves or have-nots.

Yeah, I think you're right. It's a communal goal to reach together.

You mentioned Norway there. Norway innovated in its tax system, and they seem to misunderstand that these kinds of models are variable: when you change a certain tax rate, you will see people leave. We've seen the removal of non-doms. I'm really worried about this every single day. I have friends saying, “Hey, I'm leaving. I'm leaving. Why are you staying?”

To what extent is the removal of non-doms a massive problem impacting the future of the UK?

I think this is one of those classic cases of whether you want a principled approach or a pragmatic approach. I'm a pragmatist. I do see the brain drain, I recognize it, and I do see that many of the people I know well who have chosen to leave have left. They were also incredible angel investors. They employed a bunch of people.

Do I think everyone should pay equal tax? Yes, in principle, but practically speaking, I'd rather that talent was in the UK. I'm seeing some exceptions to that. I heard about a billionaire VC who I think has moved to the UK recently. You do get some movement back in the other direction, but I would take seriously, again, leading and lagging indicators. I would take seriously the leading indicator of some of the non-doms leaving.

Yeah. One of the challenges with the UK is this tug-of-war between principles on the one side and practicality on the other. The principles have been: you remove non-dom status, change inheritance tax rules, change capital gains tax, put fees on private schools, and then assume that everybody's going to be happy to stay, really.

I mean, yeah, I just think that's too much, actually. The impulse on the system is too much, and we are shooting ourselves in the foot, really. So I agree with Tom that, in principle, as a UK taxpayer, I'd like everybody to pay the same taxes, but I recognize not everybody is at the same starting point. People do come to the country with existing wealth, really, and it can't be fully right to then seek to tax that. Therefore, there has to be some provision that makes it possible for people to stay here and so on. I think it's also part of this thing: if we're serious about building the country to be a country that clearly wants to win, then we better fix this as well, actually.

Well, this is where, for me, pandering to Trump's populism—which is like the Labour government's desire to pander to traditional left-wing policies—is destroying a pragmatic approach to wealth creation and wealth sustenance. All of the things that you said—inheritance tax, capital gains, schools—are, bluntly, pandering to traditional left-wing policy and probably don't even make economic sense. They make absolutely zero economic sense.

I mean, listen, I interviewed—I can't say it live on air, but I'll tell you afterwards—one of the most famous politicians in the country the other day. They said, "We have to get rid of the Treasury because they do not have variable models." So they literally have static models which say, "If you increase the tax rate to X, you will get Y."

Oh wow.

They do not have any variability in what happens with important export of anything. And that is why their numbers say we should do this.

Oh God.

Fascinating, huh?

Yeah, that's not good.

It's terrifying, but it worries me. Do you believe the multiplier effect? I always get the pushback whenever I'm on social. I'm like, listen, it is great having non-doms: they spend in restaurants, they hire people, they buy homes, and they spend in shops. Do you buy it, or do you think that trickle-down economics is a lie that we continuously—

I think there's a bountiful trickle-down effect, but there is also this need for fairness as well, and I think it is just a balance that we've got to strike between the two. People who don't enjoy a privileged tax status and pay full taxes are sitting in the same restaurant as people who do enjoy a privileged status. That's also not right.

We've got to find a balance between the two: how to make it feasible for people to stay here and not be penalized, but at the same time try to be as fair as possible as a country as a whole, because we need to hold hands together on this, actually, as a nation. We need both people who have come from outside the UK and people inside the UK to feel we're on a shared mission together, really. It's got to be somewhat fair at the same time.

I just think the balance right now has probably swung too far in the opposite direction, and we're actually making it much harder to do that.

I'm a strong believer in a Keynesian multiplier effect. In our small world of tech, it's the only part of the economy I know much about, and I see it on a daily basis. Like angel investing in GoCardless: if I look at some of the other angel investors in that business, they were non-doms. They were actually Europeans and some Americans.

The founders of that business built an important company for London, employing hundreds of people. One of the founders left and built Monzo, and another founder has left and is a VC at another firm in London. If you look at the number of senior talent in GoCardless that has gone on to create other businesses, it's an amazing alumni network there. It's an incredible multiplier effect.

That's what we're saying, actually. You've got to have those initial pockets of innovation and growth, and then I do think you get this real multiplier. The good news is businesses are growing faster than they ever have before, so I think those cycles will happen quicker. Previously, it might have been 5 or 10 years before you started to see the best senior operators come out and build a company. Now it might be 18 or 24 months.

Is there any change with SEIS and EIS?

Yeah, I think a lot of these EIS funds are not very effective, and VCT funds are not very effective.

Why is that? I agree with you, but I don't know why.

Because the quality of investment managers is quite low, and because they feel they've done a good job if they get anywhere close to just returning capital. Instead of saying, "Here's an investment—go swing for the fences," it's, "For God's sake, don't lose it." So you take the low-risk return and flip the company as quickly as you can. If I get 80 cents on the dollar back, I'm happy. In fact, all the returns are somewhere between 80 cents and $1.20 on the dollar. It's ridiculous. I think those funds are a freaking disaster, really.

Would you get rid of them?

I have to get rid of them. I also think there's a lot wrong with the UK tax system that is maintaining too many zombies in the UK.

Like what?

The most obvious is R&D tax credits, which is deeply unpopular for me to say. As a founder and a CEO, I'd never say this, by the way, but as somebody who's not currently a VC and who's not currently running a company, I feel free to say what I think is true.

We're currently investing about $7.5 billion a year in R&D tax credits for 55,000 companies per year in the UK. There is no quality check, if you like, on the value that's been created there. All you have to do is prove that you spent the money on something you can loosely classify as R&D, and you get a check from the government.

This is classic helicopter money. Passive money goes to good and bad. If you're going to be brutal, you'd say that either it goes to companies that don't need it, or it goes to companies that shouldn't have it.

In my view, it would be much, much, much better to take that same amount of money and put it into funds and put it into active venture. That way, when things are going well, you double down; if things are not going well, you kill it.

We do end up tying up national talent and national treasure in companies that are never going to be successful globally, that limp on from year to year living on R&D tax credits. I'd much rather see valuations go up, actually, which I know, as a VC, you're probably not very keen on hearing. I'd much rather see that because we end up with the same dilutive effect as we get this free money from the government every year.

By being actively managed, we get to recycle our limited amount of talent and our limited amount of capital into companies that are really going to make a difference. That's one thing we can do.

I say, unsurprisingly, I think tax credits are pretty important. What I hadn't thought about is that I have a sort of biased view of just higher-growth companies at the early stage of their life, where you're investing in the future. I like your point about EIS and SEIS in the same way, because I just think about angel investors. For angel investors, I think it makes sense.

But to your point, on the R&D tax credit, what I don't see is these kinds of zombie companies that have been claiming it for a decade and actually aren't necessarily building for the future. Maybe we should start to take time into account, like they do in the US with capital gains tax, and start to taper off R&D tax credits to avoid what you're describing.

Yeah, we're running at roughly 2 times the rate of the US. If you look at 4 big differences between the US and the UK, one is the retention of talent and the need to keep people in the country. The second is that the quality of mentoring at the very early stage needs to be ratcheted up a lot higher here, and I think it just needs more coordination.

The third is the excess of support in the UK for companies that are not making it, that limp on forever. The fourth is the massive shortfall in capital. I don't feel I've got quite the agreement I expected on this, but I think we just need it here, actually. We need to flood the UK with venture capital. That's what we need to do.

My takeaway from this show is that we just need to put Aatish in for the BBB lead and let him run it.

I'm not sure I'm a banker, to be honest.

I think you'd do a brilliant job. That's why you're qualified. You're literally interesting. You mentioned the mentoring there, and you said there are ways that we could do it. How do you think we could do it and increase that level of mentoring?

I agree with you. One of the things we do is, whenever we're making an investment, we will often bring in other founders from our network, people that we've worked with before. The value-add from those people, partly because they've got experience and partly because they're paying it forward, is unbelievable. I totally agree.

I always say to founders, "Never have a minimum check size for amazing angels." There are some who can only do $5,000, or some at $1,000, and you can do that with angel syndicates. That is just as valuable, and often they'll give more because it means more to them, so I really always push on that.

Obviously, we have Project Europe now, and I spend a lot of time with Kitty, the CEO.

Congrats.

Thank you.

We love it. That is very kind. I'm so pleased that you're in it.

Stan, you're not allowed.

Tom can't be.

No, he's not.

He's not stopped asking me about it for a year.

Yeah. Tom's heard it all.

My question to you is: Kitty always tells me that the biggest enemy of talent in the UK is quant funds. I was like, "I'm sorry?" And she goes, "Yep, quant funds." They go to the universities, source the best talent, and throw 250k at them straight away. The best engineering talent is just going straight to quant funds, and quant funds are much better recruiters than anyone else.

How many people work in quant funds, though? Is it a big number?

A member of our family works at a quant fund, actually, and is paid a lot of money, I think, to do something very similar. But there can't be that many people, so it can't be the biggest drain on talent.

Maybe not, but probably 1,000.

Yeah, which is 2 years' worth of full computer science and robotics graduates, which is quite a lot. I mean, 1,000 more people in the ecosystem would probably be a pretty significant needle mover.

There's definitely competition from that for the smartest quants. I think one of our jobs is to make startups even more appealing, celebrate the successes, and actually show the alternative. I think EIS, like the sort of Entrepreneurs' Relief, is a wonderful example of something that can maybe tip that balance, because often the economics from a quant fund are income tax. I think things can be done.

What would you do with Entrepreneurs' Relief to make championing entrepreneurship better?

I'd expand it.

It's limited to like £1 million or something. It was taken down.

Exactly.

For those who don't know, what is Entrepreneurs' Relief, and what does it mean?

Entrepreneurs' Relief is the ability for you to get preferential tax treatment if you've grown a company. I actually think making capital gains exempt for entrepreneurs could maybe tip that balance when you're comparing against quant funds, if that's the competition.

I do want to touch on the wider world around us in 2 ways. One is the US, and the other is China. Again, this wonderful politician that I interviewed the other day said, "You know what? We were an afterthought for the US, and now we're not even that." In a wider-world perspective, what does not even being an afterthought mean for us and what we need to do?

We do actually have, as Tom was saying, universities that are globally great universities. Cambridge is not that different from Stanford. It may be a little bit smaller and a bit less funded, but the quality of research that we're doing here is as good. So there is raw talent here.

I do think London is a really great city and a great place to live and work. It's probably the best city this side of the Atlantic and arguably the best city in the world to do this. I think it's a great place.

Do you think London's got worse? Everyone says the crime, the lack of public services, or the poor quality of public services. Do you think London will revert back to London in the '70s, which is grim, gloomy, and has no growth?

It could if we let it, but I think it's possibly not as shiny and smart as it was. I still think it's a pretty good city, and there are lots of good things to like about London.

Are you concerned that Labour will let it get to that deplorable state in the next 4 years?

I don't think they will, but I would like to see them move more quickly on policy changes and action than they're currently doing. That's certainly true. But I think they will listen and change, actually, so I'm optimistic about our ability to get change.

I think London's a special place, and I feel lucky if I compare living here to other places. It's the multiculturalism and the diversity, but actually it's just an interesting place to live. The fact that I can jump on a Lime bike, come over to do this in the afternoon, have a meeting at Number 10 shortly thereafter, and go to the European headquarters of a big brand—I could do that all on a Lime bike. It would take 5 flights to do it between those stakeholders in the US.

Actually, that proximity effect adds a real richness to life. Does it have its challenges? Yes. But there's an incredible pool of talent, so I think the petri dish for continued growth is there.

My word, if that's a standard afternoon, you're a very important person, aren't you? Jesus Christ.

I just popped down to Number 10. I popped down to a global CEO.

Only sightseeing. Wow.

I was mainly sightseeing. I just barely managed to get through the emails.

Lime Bikes is actually a portfolio company, so I'm just driving up the revenue—constantly cycling around on it.

Yeah, exactly. I'm going to expense it to your show.

Thank you so much. I'm going to get Brad to sponsor it.

That's amazing. Final one before quickfire: China is changing faster than ever. Tom, you said before, when we were walking around the block, that China is the thing you've changed your mind on.

I've changed my mind on China a lot. Strategically, I think they're in an amazing position for the obvious reason that I think more countries are open-minded to working with them, given what's happening in the world.

But I think there's a less obvious reason, and that is partly as a result of DeepSeek. More broadly, we've learned a lesson in the last 12 months: foundation models can be distilled relatively quickly. When I was on your show last time, I talked about how foundation models were going to be the fastest-depreciating assets in human history—weeks. It's almost days now.

If you live in a world where foundation models are commoditizing really quickly, then you ask where the value accrues. I think the value accrues at the application layer. We're invested in companies like Synthesia in London and Harvey in the US at the application layer.

I also think it accrues to hardware, and if I look at hardware, China is so much better than the rest of the world at manufacturing, hardware, and value-add. I think those devices are actually going to be the conduit for commoditized AI.

In that world, I've probably gone from being excited about US dominance in foundation models to some extent, to thinking that maybe the value is also going to accrue in the hardware layer. That's somewhere where I think we're playing catch-up.

I completely agree. The hardware layer is just sitting above the semiconductor layer, and I think the one that we can play in is the semiconductor layer.

I also think China is in a really good position, partly because it has made this very significant, continuous investment in startups and venture over the last 20 years. As a result, if you look at a bubble chart, if you like, of where investment is going into AI, and you color-code it for the US, China, and Europe, it's basically the US and China, with these tiny little dots for Europe. Europe is really missing.

It's the US, with China sort of chasing its tail, and I also think the geopolitics of America trying to dislocate itself from the rest of the world will put China in a much better position geopolitically as well. I think Europeans are going to be much more open to working with Chinese companies and doing business in China than they were even a year ago.

I do think things are changing, and it's probably not good for the US, but that's what I think is happening.

Do you think we should be open to doing business with them?

I do. I've sold a company to Huawei, actually. I spent about a month working for them, I have to say, because they didn't give me authority to buy a box of pencils after they bought the company, so it was interesting.

This is a country that doesn't allow our companies in there. They put their companies in ours. They acquire data on all of our consumers. We don't know where it goes. Every single piece of data that a Chinese company has, the Chinese government has authority to acquire at will.

I think all that's probably true. But they are commercial as well, so you can do business in China. When I ran this chip company, our biggest customers were in China. Our biggest customers were Huawei and ZTE.

It was easier to get them to do a deal with you and sell product to them than it was with a US or European company. You had to negotiate pretty hard on price and stuff, but nevertheless, they were willing to engage, and we built some really good relationships with them.

On a personal level, I think people are actually pretty decent, and I think that you can do business with them. The Chinese state is something different, and you've obviously got to be wary of that. But I think there's a lot of scope for us to do a lot more business in China than we're currently doing.

Do you agree?

If I look at the talent and the areas that they have decided to focus on, they're all important. Battery technology—BYD is a force to be reckoned with.

DJI is a force to be reckoned with.

You mentioned BYD. Do you not worry about the Chinese subsidization of their car industry and what it's doing to the European car markets? I mean, the German car market is being destroyed by BYD and Chinese cars. It's because the Chinese government is subsidizing between 20% and 30% of their car production.

Well, it feels a little bit unfair. Is it subsidizing BYD? Is it BYD itself?

I mean, there are variable different ranges, but it's anywhere between 8% and 25%. It's certainly cornered the market in some of the rare materials that are necessary for batteries, and I think it's got scale and the ability to compete, really.

So, in that sense, the German car industry has got other challenges. One of the other businesses I sold was to Bosch, actually, so I'm vaguely aware of what it's like working in a large German company. They have their own challenges.

Did you buy your own pencils?

Not really. I'm not the fourth acquirer, so he can buy a rubber.

I love stationery. I absolutely love that.

Listen, guys, I want to move into a quick-fire. I'll say a short statement, and you give me your immediate thoughts. Does that sound okay?

Yeah.

Okay. So, Aatish, what do you believe that most people around you disbelieve?

I think that things like R&D tax credits ought to be curtailed, and we should put the money into a lot more venture. That's a really unpopular thought, actually, but I still think it's right.

Mine would be that I keep hearing people talk about the first 1-person, billion-dollar business already being created. I think that's absolutely ridiculous. On the one hand, companies are growing faster and more efficiently than ever. Bolt now has a $40 million revenue run rate in 3 months. They're going to grow incredibly quickly, but I think we've seen distillation of foundation models. We're going to start to see distillation of business models and businesses.

I would expect these really successful businesses to get copied ridiculously quickly. I think this idea that you're going to have a sort of moat that enables 1 person to deliver $1 billion of revenue a year is a myth.

What is the distribution of value in the foundational-model landscape in 5 years?

My big one here is that I've changed my mind. I thought OpenAI was a foundation-model company. I now think it's a consumer company. It's at a $12 billion run rate or something.

My thought here would be that it's going to aggregate to the application layer, and brand is really important. They signed up 1 million ChatGPT users in an hour last week, it was announced. Brand is incredibly important. The application layer is important.

Then I think hardware, as I mentioned, is important. This is one of the reasons we invested in Nothing. We believe they've got 7 million devices out there that are potentially conduits for their AI.

Yeah, I think that might be right. The value is going to be balanced up at the application layer, but I also think there will be value at the hardware and semiconductor layer below, because it's plausible that LLMs are not the end of the story here in AI.

There are some obvious limitations to what LLMs are going to be able to do, so there's more innovation to come. That's going to change the models, it's going to change the math that we have to do, and so on.

Some of the things that are going to be constant—we're still going to be doing very large matrix-vector multipliers at high speed in silicon. I think that's the sort of thing where we can build a competitive, long-term advantage. I think there would be value accruing—perhaps even more value accruing—to companies like NVIDIA, which I think will be big. At the application layer, exactly as you say, I think there'll be value accruing there.

How about inference at the edge as well? That's something you understand better than me.

Yeah, these models are lighter, so more and more could happen on-device.

Yeah, that's true. But with that is coming a lot more chain-of-thought reasoning and a lot more test-time compute. The token generation is still going up, actually. I still think there's going to be a large amount of silicon required to do high-performance inference, even at the edge. So, there's a lot of scope in inference.

You know what I just can't get? I can't get how, if we all appreciate the shift in focus from training to inference, Jensen and NVIDIA are just sitting there going, “We're going to get screwed because our architecture means that we're not optimized for inference.” That is not happening. Jensen is not just saying, “Fine, we'll just enjoy the training era while it lasts.”

So, help me understand: where am I missing this?

They're making a bunch of architectural changes to GPUs to make them better and better for inference. There is a lot of architectural change going on there. It obviously wouldn't be a big surprise if we saw Jensen starting to adopt and reinvent himself as an in-memory-compute company. That wouldn't really surprise me; he'd be working on that.

Whether he does that organically and internally or through some sort of acquisition remains to be seen. But I think it's certainly likely that he's got the resources and the cash to move the organization, or build an organization, in pretty much any area he wants.

One thing about Jensen—I spent about 1.5 years working for him—is that he definitely pays attention to and listens to the market. He's got very big ears and tracks what's happening with enormous scrutiny. I do think we've got to expect NVIDIA to be tracking in the direction of becoming more efficient at inference.

What's your biggest takeaway from working with Jensen?

Firstly, he's a good human, so that's good. One of the world's richest people is actually, I think, a good person.

He is, however, a bit of a control freak. Many a time, we're just about to give a presentation to a major customer, and Jensen wants to go through the deck. We'll change the product name, schedule, pricing, resources, and everything on the fly, with 10 minutes to spare before the meeting.

He's quite hard to work for in terms of his desire to have command of detail and to be in control of the most important variables in the company. But, in a way, as a founder, I do respect that.

Within NVIDIA, we used to have Jensen at the top, and we had a layer of people whose job was to buffer everybody else in the company. This buffer layer would deal with Jensen, which was great.

Human shield.

A human shield. There were people below that who could actually get on with stuff.

I obviously like the guy, and he's an incredible communicator.

We always hear about his direct reports. I don't have direct experience with them, but he has so many—like 50 or 60—and it sounds great when you hear him talk about the 50 or 60. We never hear from them. Is it good for them?

Well, it's a brutal culture, I'd say.

Yeah, it is.

But in a way that is not malevolent, if it's possible to imagine that. He will tear people apart in public over things that they haven't got command of or that he thinks they're wrong about. He will rip them to shreds and leave them whimpering in the corner to lick their wounds.

But I think he then forgets it and hopes that the exercise will have resulted in some improvement in the way the person thinks and acts. It's not for everybody, that style of management, but honestly, you've got to admit it's worked. He has done an amazing job—an unbelievable job.

Would you buy OpenAI at $300 billion? Yes or no? Why?

Yes. If you look at it as a consumer business, I think it has extraordinary momentum, and it's only just started integrating moats. Historically, there have been no switching costs, one of the most important moats of a business. But now people have started using it, and I actually think the memory is helping people stick.

If I look anecdotally at my kids at school, for them, LLMs are ChatGPT. They're very well placed. Do I think the same thing about my mother on the other end of the age spectrum? Yeah. They own brands, so it's incredibly powerful.

It doesn't mean I think it's the best GenAI investment, but if I were sitting independently, do I think it's a good investment? When your downside is somewhat protected and they're at a $12 billion run rate—let's say it's a 20-times forward multiple—I think it's a reasonable place to put money.

Yeah. I think a lot of the demand on these foundation models is going to be through APIs, used by application software. Those APIs are going to be driven by latency, performance of the model, and so on. Things like Claude are as good as, if not better than, OpenAI's models.

Given that agents are going to be calling APIs, that demand will be driving a lot of demand here. It's not obvious to me that the consumer chat interface is the winning interface. It seems to me that the API interface, with applications calling agents, might be a bigger interface. So, I'd probably put the money elsewhere.

Both are good answers. Would you?

I would.

Yeah, I would. I always love businesses where everyone thinks they're reaching the top, and then actually they're just reaching escape velocity.

I think the same with Revolut right now. People think a $45 billion or $60 billion new round is pricey. I would buy the hell out of Revolut right now. But I totally agree with you in terms of introducing the moats and the memory. I think the memory is so important. You go back and they remember what you did in the past. I'm always doing past searches; I'm always going back there.

And actually, I do. It's so funny. For every single show, I put the prompt into Grok, Perplexity, and OpenAI: “You can buy and hold 1 public stock for 10 years. Which one do you pick?”

I mean, I'm so concentrated in tech. I'll avoid tech stocks and say a uranium ETF. I have concerns about the costs of energy for productivity. I think climate change is real. I think the best source of energy going forward is nuclear fusion and potentially fission, and SMRs are going to be important.

I think it's the most predictable, cleanest energy source we have. I'm not betting on 1 individual company; that's difficult to do. So, I think if I take an ETF in uranium, I might enjoy the upside of the market because it'll be needed.

Love that. What's yours, Dad?

Yeah, I probably would avoid tech as well, actually, for the same reason. Probably Rolls-Royce, actually, because I do think defense is going to be a big kicker in terms of demand.

So, the aero-engine business—and I mean, it has actually gone like a rocket this year anyway, that stock. I mean, it's gone like 3x this year. But I actually think we're at the beginning of a journey, and I think it could be much bigger because, as a European aero-engine vendor, I think it's going to see high demand, actually.

Yeah, and you might catch fission as well.

You might—fusion.

Fusion.

Yeah. No, fusion.

You can snap your fingers and change 1 thing about the UK tech ecosystem. What would you change?

Well, I think flood it with venture capital. Seriously, I think that's the 1 lever we can pull that will make a big difference. Everything else will take time and stuff, but I do think a lot flows from capital availability.

I love that. I would say sentiment at the moment. I think there's more—this question is being asked so much that it becomes a drag. What's the most underinvested but exciting area today?

Yeah, you, Tom—you do this for a living. I'll go hardware.

I'd go 1 level below semis. I think semiconductors fit into the hardware that Tom's talking about.

Which politician do you most respect and admire, and why? Given your specialization?

Yeah. So, I'm going to stick to the UK. At the moment, no one in the current government really fills me with enormous enthusiasm. I think Patrick Vallance is a useful guy who's trying his best to make an impact on the UK.

But he's not really a politician; he's doing the job of a politician right now. I do think, in the current government, Darren Jones has the potential to be great.

Final one, guys: 10 years' time, where is the UK, and how many $10 billion companies will we have on the LSE then?

Yeah, I think we will get the UK pointed in the right direction. I think it will require some government embracing of the challenge and a lot more communication by government on what we're going to do and how we're going to do it.

But I think we're approaching a point. We're about a year into this current government, with 4 years to go to the next election. Things have not gone well, I think it's fair to say, and I think we're approaching a point when they've got to recognize a change and make some changes.

I think we are going to see some changes that will be positive, and assuming that happens, I think in 10 years' time we will have achieved this $500 billion valuation in tech. The UK will be seen as the magnet in Europe in which people come to build these companies. So, that's what I think we're going to achieve. I'm an optimist.

I think sometimes the best companies grow from adversity. I think partly because of the concentration of talent, they'll just aggregate more than they have.

So, I don't know—1999, was it Salesforce? And then you have Airbnb and Uber in 2008. I think we'll look back, and the companies that are most impactful in the decade will have grown in the UK. They won't be names we know today because these companies are growing faster than ever.

So, they'll be AI-native, incredibly fast-growing businesses, and it's not clear to me they'll list at all. If you look at the trend direction there, we spent a lot of time assuming listing makes sense, but some of our best portfolio companies, like Stripe, aren't listing anytime soon, and they're finding ways to deliver liquidity.

So, I wonder whether we'll even be talking about whether they did or didn't list in the UK.

God, that's opening up a can of worms. I mean, you could spend another 2 hours on that, but I cannot thank you both enough for joining me. It's been such a fantastic discussion. Honestly, you were the 2 people I most wanted on YouTube because I think it's such a different perspective you both bring. So, thank you so much for doing it.

Thanks for inviting us. We've enjoyed it.

Loved it. Really fun.

Thanks a lot. Amazing, guys. Thank you so much.