Europe has startups. Why not more tech giants?

0:01Everyone, I finally entered the 21st century and it's not because of these new microphone covers we've got, by the way, if you're watching the tech download ones, it's because I've got brand new internet. So, I don't know, I was getting like 100 megabyte download speeds before, I'm now getting a gigabyte download speed, wirelessly, 2.5 gigabytes, allegedly, if I plug in my ethernet cable. This is huge. This is huge. This means my calls won't drop out. And more importantly, it means I'm ready for GTA 6 when it comes out this year. That will just be downloaded so quick.

0:36Hello and welcome to the Tech Download. I'm Arjun Karpul, CNBC's senior technology correspondent. Now, before I tell you who our guest is for this episode, I want to tell you a little bit about him. He's one of the most influential people in European tech. In the 1970s, he co-founded a company called Acorn Computers. Now, this was really important because within Acorn Computers, the first ever ARM chip was designed. It was an innovation at the time, a big breakthrough in chip technology, offering low power, high performance computing for the computers that Acorn Computers were producing. Now, ARM was eventually spun out of Acorn Computers.

1:13Now, for those of you who haven't heard of Arm, it is one of the most important semiconductor companies in the world. They effectively create a blueprint that many chips are based on. Pretty much every chip in the world that's in a smartphone at the moment is based on Arm technology as well. And it's got a fascinating history. It was founded in the UK and it was sold to SoftBank for$32 billion in 2016. That company now, Arm, is worth more than$200 billion. So a big success story out of Europe. Now, my guest has a career that spans nearly five decades. And he is, of course, an entrepreneur, but also an investor.

1:54He's the founder of Amadeus Capital, which was established in 1997. This is a venture capital firm. And right now he's interested in what comes next in AI, in quantum computing and in next generation compute, such as new architectures around semiconductors as well. He believes Europe can compete in this new wave of innovation in areas like quantum and AI. His name is Hermann Hauser.

2:23You've lived through PC, Internet, smartphone revolutions. We're now being told this is the AI revolution. What is genuinely different right now about AI and what looks like familiar hype through the previous cycles you've been to? First of all, I must tell you that before all these waves that I got involved in, I was involved in AI because I almost changed my PhD from physics to AI because there was so much excitement about AI for a short term. Fortunately, I didn't change my PhD to AI because just after that initial excitement, when we thought we would have a world chess champion with our clever approach to playing chess in half a year, well, it took another 15 years.

3:10So this wave of AI coming and going, you know, is something that hasn't just started. But now your question was, what's different this time? Well, the results. You know, the results are coming in in terms of dramatic productivity increases, especially in software programming and also customer relationships. You know, it's a very clear advantage to companies that use AI. It's not just hype. So from an investment perspective, how are you investing? Because when you look at the spectrum right now, a lot of the value is accruing at, let's say, the picks and shovel layer. And the large valuations are being seen at, say, the frontier model layer.

3:57And there's this view the frontier models are going to come after everything right now. So when you look at the landscape as it stands, how do you think about investing? Where do you see the opportunities? I always make the analogy with electricity because there are many similarities between the rollout that we saw some years ago after the first industrial revolution when steam was replaced by electricity and AI now. The first analogy, of course, is that it is really general purpose. It is a technology that pervades all of the economy, all of the different sectors. And if you look at the real impact of electricity and how electricity generated an enormous amount of value for our national economies, it really was the application of the generators and the motors in each of the different sectors.

4:55And I think that's what we're seeing now. Of course, there are some sectors that are more easily adapted to AI than others. And the one that's producing spectacular results at the moment is actually coding. I just talked to, you know, the head of Microsoft Research in Cambridge that I have had a hand in setting up many years ago with Bill Gates and Nathan Mervel. And I said, how much are you using AI for your researchers? And he says, Herman, it's unbelievable. We've got researchers that have got 150 agents running overnight, you know, iterating things. So a single researcher has a whole, basically a whole lab coding for them.

5:41Then in terms of where you're thinking about investing and how you're thinking about investing into the AI theme, where are the opportunities as you see them? So it's really applications of AIs in different sectors. Call centers, for example, because voice recognition is so good now. The question is not, can you understand what people say? But it's dialogue management. How do you respond nicely and in an informative way to people's questions? And we've got an investment in a company called PolyAI. It's again a spin out from the engineering lab in Cambridge that is now doing extremely well in the US.

6:25You know, having FedEx as one of the customers. So this is something that has real rollouts in really large companies all over the globe. What does this all mean for valuations these days? And when you look at some of the AI companies, are today's AI valuations justified? Well, again, the analogy both with the Industrial Revolution and, of course, more recently, the Internet, I think, is justified. So during the internet bubble, things got a little bit ahead of themselves and some of the phenomenal valuations then collapsed. And my expectation is that the same thing will be true in AI, that because of the productivity increases that we already have seen in some areas, and I do expect these productivity increases to be rolled out across all different sectors of the economy.

7:20this is a revolution that will create more value than probably any other technology revolution that we've ever seen. That the road there is going to be a bit of a roller coaster is also something that I expect. And some of the valuations have clearly gotten made of themselves. And some of them, of course, based on these gigantic circular orders that are being placed between NVIDIA, OpenAI, Anthropic, and so on, in the hundreds of billions. And if you do the math, that doesn't seem to work out. And does that concern you at all, Herman, in terms of the potential ripple effects, if there is any kind of a rupture in that chain?

8:03I think there will be ruptures and there will be reassessments of these valuations. But if you look at all these really large companies like OpenAI and Anthropic, they're not going to go bust. They are very stable. They typically have large cash cushions so they can stand a disruption for a while. So I think it will be hard on some of the investors in particular because, you know, some valuations will go down quite a bit. But I think overall, the sector will prosper long term.

8:44Executive Decisions is the new podcast from CNBC, where I ask powerful leaders about their decisions that changed everything. I'm Steve Sedgwick. Here's the CEO of the London Stock Exchange Group, Julia Hoggett. There is no justification whatsoever for not doing the right thing, even if the right thing is hard to do, which also means that if the context changes or the information changes, it is absolutely fine to also change your mind. That's Executive Decisions with me, Steve Sedgwick. Get it wherever you're listening to this.

9:37it's fair to say Europe lost out on that last wave of innovation around the internet, around social media, around internet services, and around cloud computing. I mean, an example is the fact that the majority of the European cloud computing market is basically dominated by Amazon, Microsoft, and Google, all three American companies. And so as you think about this sort of AI wave, this quantum computing wave, there are a lot of debate right now as to whether Europe can compete, create huge global companies that are able to challenge those companies and services and technology coming out of the US and China.

10:17One example is Mistral AI. Now, we had the CEO, Arthur Mench, on a few episodes back on the Tech Download. We had a whole conversation about what they're up to and their position within the broader AI ecosystem. So many see this as kind of Europe's hope to compete with the likes of OpenAI and Anthropic. And there's lots more examples across the board, whether that's in semiconductor space, whether that's in the model front and the application layer. This is a big debate happening right now as to what Europe's position is going to be in this new era of technology.

10:51Is AI and the quantum computing era, which we'll certainly talk more about, any different? And if so, why? Universities in Europe have an IP output that's comparable to the US. So we don't have a university problem in Europe. We, in fact, arguably in some area, and quantum is one of them, we produce more high quality quantum results than American universities. If you then look at the next stage, which is the startups and spin outs, it's not a widely known fact, but it is a fact that Europe produces more startups than the US. So we don't actually have a startup problem. But then we've got a real problem with scale-ups.

11:33And that is the reason why the European Innovation Council, which I'm on the board of, has just started this scale-up Europe fund. And we've just selected EQT, the second largest private equity company in the world, as the manager. They have got$300 billion under management. The first EIC fund, the European Innovation Council, is 10 billion. It's the largest VC fund in the world for early stage. And that's going very well. For every euro that we invest, we get four euros from the market. So it's a catalytic fund and it's not crowding out the market. It's crowding in the market. But with the Scale Up Europe fund, we've chosen a different organization where we just put in$1 billion as the cornerstone investor of a$5 billion fund, where 80 % of the money actually comes from the market.

12:26So I think that move towards very much a market-based approach to solving our scale-up problem is in order, and I have high hopes that we'll be able to keep some of our great startup companies that often migrate to the US and Europe. Out of the 12 unicorns that I've been involved in, every single one of them ended up in the US. There's nothing wrong with the US, but, you know, we ought to keep some of them in Europe as well. When you talk about scale ups, we're talking about companies here from startup, you know, to billion dollars to 10 billion to 100 billion, you know, these massive global kind of players.

13:01Is the challenge money, though? Because I've seen more and more over the past few years that, look, we've had big VC funds globally come into this market. Even companies like SoftBank with the Vision Fund come into Europe. So is it less about money and more an ambition question? Because one of the things that still happens, and we see this a lot, is either they migrate to the U.S. or U.S. companies just come and buy out these very big players. And we saw that. Look, we think back to Google DeepMind back in 2014 when that was bought out for a fraction of the valuation it's worth. Now, you even think about Arm, which, of course, was founded in Acorn Computers and spun out the company Yuko founded.

13:42that was bought by SoftBank I think at the time in a deal worth roughly 32 billion euros the valuation of that has absolutely soared. You know Julia Hoggart the CEO of the London Stock Exchange tells the Arm story it's her story not my story we made three mistakes with Arm the first mistake was that we sold it to SoftBank for 32 billion it was a 43 percent premium of the price that Arm was quoted at on the London Stock Exchange and Nasdaq it was a dual listing on the day. And of course, all the fund managers in London headed for the exit. The second mistake was that we didn't manage to get it listed in London again.

14:22And it was just one minor regulation that Sonsan objected to, which, by the way, Julia managed to change six months later. So now, if it had happened, maybe we would have a dual listing and I hadn't given up on the dual listing. The third mistake was that Arm went back to all the fund managers in London who made lots of money with Arm and said, look, you have a chance of buying back in, albeit at 50 billion. And they all said, 50 billion? You must be joking. Which European company have you ever come across that grew above 50 billion in technology? Well, it's now around 300 billion. So they missed out on the biggest creation of value in the UK, actually in Europe, ever in a three-year period.

15:04Arm is now the largest company in the UK by a fair margin. It took the German car industry, which is Mercedes, BMW, Volkswagen, and all the suppliers, 100 years to create an industry worth 200 billion. Arm created 250 billion in three years. So that just brings home the importance of technology for the growth rate in a nation. If you're not in technology, you're just not growing your economy. So just coming back to that question then in terms of what is missing for Europe, is it just money or are there other parts in order to have successful scale-ups? No, but let's start with the money because without the money, you're never going to get there.

15:47It's one of the necessary ingredients. And with the Scale-Up Europe Fund, we're very much trying to address that. The second problem after that, of course, is the European Nasdaq. If you look at all our successful technology companies, they naturally go to an aspect quote rather than a European quote, although the London Stock Exchange is making good efforts to be a technology stock exchange as well. The single most difficult problem to solve, however, is a management problem. What the greatest advantage of Silicon Valley is, if I want to grow one of our companies, as we often do in Silicon Valley, from 100 million to a billion, then you go to the guy at Google or Meta or Apple who runs a billion-dollar division, and you say, would you like to go into the scale-up?

16:42We've got a great share deal for you. And if he or she likes it, two weeks later, he is the CEO of that startup. So contrast Europe. First of all, with somebody who runs a billion-dollar division of Mercedes or Rolls-Royce, look at it at all from a cultural point of view. Secondly, if he looks at it and he likes it, he is now locked up for six months or a year, typically a year, or has a year notice period. So by the time he can join, the opportunity has gone. And thirdly, the option schemes, although that has gotten a lot better as well, we couldn't give the option packages that are customary for VC-based companies in the US because of the laws in different European countries.

17:31I'm pleased to say the UK is an Anglo-Saxon exception. But later this year, we will introduce EU Inc., which is basically the EU version of a Delaware corporation, where all the share structures, the voting rights, the option schemes are like in the States. You've spoken in the past about Europe's reliance on non-European companies for key technology. This idea of sovereignty is rising. How do you define it? What role does it have to play? It's an absolutely key concept and also hits the headlines all the time. Sovereignty has to be redefined. You can't define it as doing everything yourself, because even the US and China, who are two big competitors, of course, have no chance of being sovereign in the sense of doing everything themselves.

18:23So then the question is, how do you have a fair relationship, a fair trade of critical technologies across the nations? And each nation has to ask itself three critical questions, because First, not having these critical technologies is disastrous. You have to have them. I mean, a 16-year-old will probably think his life isn't worth living if he doesn't have a smartphone. So you need it to keep your nation happy, but you also need to have all these critical technologies to be able to govern a nation. The second question is to have access to these critical technologies from a number of independent countries so that no single country can call me to ransom by withholding the technology, as indeed Trump has done.

19:08If there is a no to the second question, the third question is, do I have unfettered, long-term, longer than five years, guaranteed access from a monopoly or oligopoly from a single country? And if you get three no's, it's one of those situations where you have to do whatever it takes, unless you're willing to be exposed to this economic coercion that is associated with being dependent on a critical technology from a single country. So when you look at Europe's position now, is there a single dependency right now that concerns you most? Well, there are a number of dependencies. One, of course, is access to electronic design software from a US duopoly.

19:53That's Cadence and Synopsys. You know, we, of course, have a monopoly position with ASML in the supply of lithography machine. Nobody can make state-of-the-art chips below five nanometers without having an ASML machine. And then there is the supply of rare earth minerals and others from China. So there are lots of these dependencies. And the question is, how can you create a working global relationship in this new nations interacting, which is different from the global order that we used to have pre-Trump? Yeah. Do you think that's here to stay, Herman? Because we had over the past few years, both through the current Trump administration, the previous Biden administration and the first Trump administration, this continued and ramping up of export restrictions and various other restrictions, particularly on China, aimed at the semiconductor sector, now filtering through to AI, etc.

20:54And we see that continue. But there was always this feeling that, oh, no, it wouldn't happen to Europe. You know, it's all fine. And then we saw earlier this year when the U.S. pulled access to Anthropics mythos model to countries around the world, including many here in Europe as well. So was that moment a kind of wake up call to some extent? And as it stands, you know, could Europe survive a rupture on a technological level with the US the way that it's happening with China? Well, it's clear that there's got to be a new way of doing it. Europe has to have its own access to a foundation model.

21:33Fortunately, we have a very credible company with Mistral producing a European foundation model, which is not quite up there with OpenAI and Anthropic, but not far behind. And of course, we've seen ASML, for example, putting a lot of money into Mistral recently. And I think there is another major round of funding coming up for mistrial. So there has to be a European option, of course, unless we want to be completely dependent on the U.S. Of course, it's very important that Europe retains a very close relationship with the U.S., but it shouldn't become a technology colony of the U.S.

22:19look i don't profess to be an expert in quantum computing it's incredibly challenging and difficult to understand but but we're learning as we go and it is a fascinating area but let's just do something fun i just asked one of the ais to explain quantum computing to me like i'm five um here's what it came up with quantum computing is a super fast new kind of computer that uses tiny magic rules of nature to try every possible answer at the same time instead of checking one by one. Now look, we laugh, but it's a really interesting point and a very important point. Normal computers or classical computers store information in bits.

22:59Each bit is either a one or a zero. Think of it like an on and off switch. Now quantum computers use quantum bits or qubits which can be zero or one or something in between this is really important for its ability to carry out complex tasks that classical computers can't now the promise of quantum computers is is interesting many have promised pretty lofty results from quantum computers for example drug discovery now you need quantum computers to be able to assess kind of the makeup of molecules, how they react to other molecules. And this could lead to things like new healthcare initiatives, drug discovery, etc.

23:39So these are some of the sort of big goals that the quantum industry is setting itself as well. Now, the other point here is encryption. Right now, we have encryption on so many of the different services we use. Think about your banking app or some of the messaging apps you use as well. Encryption is used to defend against bad actors, against hacking in the form of messaging. It can be used so the company providing those messages services can't see your messages and hackers can't get to your messages as well. However, many have said that when at a certain point, quantum computing is going to be able to crack that encryption, therefore rendering modern cybersecurity pretty defenseless against new quantum computing.

24:24That's a big problem, and that's why the quantum computing industry is trying to figure out what comes next, and they're looking at new forms of cryptography in the quantum era as well. But the other point here, as much as there's so much hype around quantum computing, quantum computing is still not seen as commercially useful, i.e. it's not yet able to do things that classical computers can't. And that's also important as well as we think about the future of this technology. Still, it has garnered huge investment. And that's been an interesting thing as investors bet on quantum computing to be one of the next big technology waves.

25:02Just a quick stat for you. In 2025, quantum technology startups reached$12.6 billion in funding. That is 6.3 times higher than in 2024, according to McKinsey and Company. So it just shows you the amount of capital and excitement that is happening around this technology right now.

25:25Herman let's get back to some of the the big themes you're investing in I want to just start with sort of I guess a broad umbrella turn of next generation compute semiconductors and what comes next because so far this AI boom has been built so much on a handful of players you think about Nvidia and their graphics processing units AMD the memory chips coming from South Korea etc. You had a critical role in the creation of Arm, the company whose designs are basically in nearly every smartphone in the world and other consumer electronics. So you have a great, great view into this as well. Arm was almost born out of the first chip design was almost born out of scarcity, right?

26:08Because you needed a solution that didn't quite exist yet. Are there any parallels happening now in sort of the semiconductor space that you saw with the AI boom that you saw back then with the creation of the first ARM chip. And what does that mean for kind of where you're thinking about investments? Yes, very much so. The creation of ARM was based on a technical breakthrough that actually came from John Hennessy at Stanford and David Patterson at Berkeley. And it's one of the very few examples where the invention was made in the US, but the commercialization was done in Europe. with ARM. So there was a technical hiatus, a breakthrough with reduced instruction set computers, which showed that less is more, that complex instruction set computers actually are bigger, more power hungry, and less performant than the smaller reduced instruction set computer that consumes less power and produces higher performance.

27:16A similar change in technology is happening right now because of the energy consumption that everybody is getting very worried about. All our computer architectures are based on von Neumann machines. And von Neumann originally laid out a computer architecture that separated the memory from the processor. And if you look at where the energy goes, 99 % of the energy goes into transferring the data from the memory here to the processor. The processor is a little bit of work and then writes the data back to the memory. And the brain does not do that. If you look at a neuron, a neuron has about 1 ,000 or 10 ,000 connections.

28:00Note that the data that are basically encoded in the synapses of the brain never moves. So you have this new idea of in-memory compute that is a new silicon structure that reflects the brain architecture much more closely than a von Neumann machine. That allows a reduction of two orders of magnitude in power consumption of the computer. So this is well worth having. And so there are a number of startups that have already invested in four of them, including Fractal in the UK and Semron in Dresden, who are silicon implementations of this new in-memory idea. I never thought that we'd have a very fundamental change in the computer architecture as a result of AI, but because of the high energy consumption of AI, this is happening.

28:54Now, in addition to the change of the traditional semiconductor industry, we also have the rise of optical computing, which uses photons instead of electrons, which are much lower power, again, but one or two order of magnitudes. And then the real solution in terms of performance, of course, are quantum computers, which are exponentially faster than classical computers and also much lower power. And Herman, this is being driven by, I think, some of the factors you mentioned there, but also, I guess the fact that there are so many bottlenecks in the supply chain. And our viewers and listeners will know there is a huge shortage of memory, chips in particular, and a massive price rise.