In short
EUVC Podcast Episode Notes: E551 - This Week in European Tech with Dan, Mads & Joe Knowles
Podcast Title: EUVC Episode Title: E551 | This Week in European Tech with Dan, Mads & Joe Knowles Co-Hosts: Dan Bowyer, Mads Jensen Guest: Joe Knowles from Smedvig Ventures Release Date: [Date Not Provided in Transcript]
Episode Overview In this episode, the hosts and their guest delve into the current state of the European venture capital landscape, comparing it to the US, discussing mergers and acquisitions, the importance of exits, and strategic investment in sectors like defense technology and energy. They also explore the implications of AI developments, including the much-anticipated GPT-5.
Key Topics Covered
- US vs Europe in Wealth Creation
- Key Insights:
- A report highlights that while the US has a significant lead in wealth creation, Europe has been growing its number of startups and capital efficiency.
- Statistics Discussed:
- 40% of the world’s top 100 unicorns are based outside the US.
- The number of startups in Europe increased from 8,000 to 35,000 over the last decade.
- The median GDP per capita in the US is almost double that of Europe.
- Late-Stage Funding Gap
- Discussion on why late-stage funding is a challenge for European startups.
- Pensions and IPO market dynamics were identified as key barriers.
- Mergers and Acquisitions (M&A)
- A record high of nearly $100 billion in M&A activity was reported.
- Significant deals, such as Google’s $32 billion acquisition of Wiz, were highlighted.
- The importance of M&A in recycling capital within the venture ecosystem was emphasized.
- Defense Technology Investment
- Discussion on the cultural shift in Germany regarding defense investments, with Porsche and Deutsche Telekom initiating a €500 million defense fund.
- The panel highlighted the importance of dual-use technologies and the changing landscape of European defense investment.
- Energy Innovations
- Emerging technologies in energy production and storage, including advancements in lithium recycling and the potential of sodium-ion batteries.
- The high cost of energy in Europe compared to the US was noted as a significant barrier.
- AI Developments
- GPT-5 Launch:
- The new AI model aims to unify reasoning and speed, but initial feedback includes some drawbacks in user experience.
- The impact of AI on compute power needs and energy costs was discussed, emphasizing the urgency for Europe to enhance its power infrastructure.
- Chip Industry Dynamics
- Discussions on NVIDIA tariffs and their implications for the global chip market.
- The need for a Chips Act 2.0 in Europe to foster local semiconductor production was underscored.
- Spotlight on Italy's Startup Ecosystem
- Italy achieved a record €655 million in startup funding in the first half of the year, reflecting a vibrant entrepreneurial environment.
Conclusion The episode wraps up with reflections on the need for continued investment and strategic shifts in Europe to enhance its competitiveness in various tech sectors, particularly in light of global dynamics in AI and venture capital. The hosts express optimism about the future, with a focus on collaboration and innovative solutions to existing challenges.
Key Takeaways
- Europe is gaining momentum in venture capital, but faces challenges in late-stage funding and exit markets.
- M&A activity is crucial for capital recycling within the ecosystem.
- Emerging technologies in energy and defense are key areas for future investment.
- Italy's startup scene shows promising growth, demonstrating the potential for European innovation.
Next Steps
- Follow EUVC: Stay updated on European VC trends and insights by following the EUVC podcast at [eu.vc](https://eu.vc).
---
This Markdown file summarizes the podcast episode's discussions and highlights key concepts, arguments, and notable points, making it accessible for readers interested in European venture capital and technology trends.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Transcript
Automatic transcript. May contain errors.0:00Welcome to Upside, where we dig into the real stories that live behind the headlines affecting European venture. Today, it is Mads, myself, and we have a very special guest, Joe Nault.
0:22This show is not investment advice, and the hosts of this episode may be invested in the funds and companies featured. Now, Joe is from Smedvig Ventures, who are a London-based VC, leading A and B rounds across Europe. He's a B2B guy like us, investing in software and software-enabled businesses, a very super-duper outdoorsy, keen skier, surfer, and all-round cool dude. Welcome, Joe. Thank you so much for joining us. So we've got a bit of a left-field-y conversation about the history of US wealth creation. M &A is up. What does this mean? Mergers and acquisitions. Lots of acquisitions going on.
0:57We've got German Megafund. We've got another defense fund that's come out of the woodwork. So we're going to talk about that. Energy. Now, I've seen there's lots of news around energy, nuclear generation, battery storage, battery technologies. And then last up, we've got the latest in AI. As always, lots and lots of stuff going on in AI. GPT-5, taking a look at that. Some big deals, deep seeks, chip needs, and a few more topics. Let's crack on, gentlemen. Is the US the only place to make money? I was reading Ruchir Sharma. He wrote an interesting piece this week, challenging the widespread belief that the United States hold this unique and unassailable position in wealth creation.
1:36Now, around the world, there are many organizations of people who have built remarkable returns in wealth, as we all know. But to break down some of the figures, he looks at what he calls compound performers or companies that are returning more than 15 % annually. Now, of the 444 with market caps over$10 billion, 248 are outside the US. Of the estimated 13 trillion in private assets, nearly half is held outside the US. Now for VC and infrastructure, it's slightly more. It's about 50 % or just over. 40 of the top 100 unicorns are outside the US. 70 % of new billionaires are not in the US. However, only 2 % of the world's top 50 tech companies by market cap are European.
2:15Europe has only around 50 of his compounder firms. That's about 20 % of the market. it. But if we look at VC, it's growing five times faster than our lovely cousins over in the States. And the number of startups in Europe has grown. The numbers seem to be a bit blurry depending on which data set you look at, but it looks like from roughly 8 ,000 to 35 ,000 in the last decade. So growing and growing and doing fabulous stuff. So US obviously is a machine. We all know this, but they're also great at their own PR. So a question for the room, and maybe Mads, I'll pick on you first. What's wrong with this story?
2:51I think it's a very generous interpretation of what the US is or isn't doing. It's almost sort of presented as if the US only being 44 % of the global compounders is not quite the achievement one would expect. But they're putting into context of the United States being sort of what, you know, less than 5 % of the world's population and still creating, you know, more than almost half of the world's compounders. That's one thing. And I think the second thing is going to look at this in the context that this is the number of companies without the market cap. And we all know if we look at the market cap, the concentration to the U.S.
3:24will be even bigger. And that's one of the reasons why today, shockingly, over the last 15 years, we've gone from a Europe that looked like it was relatively close to the U.S. performance wise to the U.S. today having a median GDP per capita. That's almost twice Europe's, if not more. So, I mean, there is just more wealth that's been accruing there. The question is, you know, how can you unpack some of that stuff? There's nothing happening anywhere else. And I think there are some green shoots that he's pointing out in the article you referenced. So Europe has got more than 50 of these global compounders.
3:55It's 11 percent. It's not 44 percent. But we also have a lot less capital going into building these businesses. And that's one of the reasons why the internal rate of return, when you look at venture capital, actually is a little bit higher for European VC than for US VC, according to Cambridge Associates. So, you know, why is Europe still behind? I mean, there are all kinds of things there around capital allocation and how much money is invested and how do we build these capital accumulation and get enough money to the right companies. And I'm sure we're going to unpack that. But I would love Joe's thoughts on some of all this before we go there.
4:29Yeah, go for it, Joe. I mean, you invest across Europe. You've been in the game for a long time. Not saying that you're old, my dear man. The other thing I will very quickly point out before I hand over to you, Joe, is that Richie is Indian and there were lots of references to India. And it felt like there was a slight skewing to the motherland for him, which I don't blame him at all. But it did, to your point, Mads, it did feel like there was some skewing of the content and the article just to make it sound a little bit more non-US and a little bit more India and a little bit more other, which I'm not convinced by.
5:03But Joe, what would you pick up on here? I agree with what Mads said, but I think the US has undeniably been a huge value creation machine, and that's been great for the US, and it's actually been great for everyone else as well. I think an important piece of context is, to Mads' market cap point, the numbers are heavily skewed by the fact that the US won the internet race. And so there's sort of five to 10 internet infra hyperscalers that have just had completely outsized value creation. And that's obviously great for the US. Would have been lovely if that happened in Europe. But it's very concentrated, right?
5:36It's a small number of companies and individuals that have benefited from that. I think that does skew the overall perception. And actually, if you strip those things out, there is a slightly more balanced picture. I agree with Mads, though. It's still the sort of comparison benchmark of, oh, the US is only 15%. It's still a slightly skewed way of viewing it. But I do think you've got to bear in mind that context of those very small number of hyperscalers that skew the numbers. I think if you look at Europe, though, what have we got? We've got some amazing engineering talent. We've got entrepreneurial spirits, amazing entrepreneurs, and actually a culture of founding businesses.
6:10The stat you referred to, Dan, around the rate at which it's growing and closing the gap in the US is very promising. We've got great early stage funding, including some smart government incentives. I think the issue really is around later stage funding and the exit market, which is in part a byproduct of this kind of ecosystem that's emerged around the hyperscale. I think that's probably the biggest issue we face in Europe to this point. There's some other things going forward that we'll come on to around AI and choking points and power of compute. But I think just in the rear view mirror, it's that late stage funding and exit market that I think has really held Europe back.
6:44I'm sure we can discuss a bit more why that is. Yeah, no, I agree. I mean, just for context, it looks like the US has roughly double the number of startups in action at the moment than Europe and the UK, about 73 odd thousands. And it is, as we've been discussing, there is a massive growth rate difference. Europe is growing much, much quicker than the States because there's obviously a long space to catch up. Focusing specifically on Europe, what levers do we have and what competitive advantage and where should we lean in? European companies generally are more capital efficient. The cost to create an additional dollar of revenue for a European B2B startup is lower.
7:27It costs, I think, more than two times as much to create an incremental dollar of revenue for a U.S. B2B company as it does for a European company. So, you know, we're just more capital efficient. And, you know, part of that is sort of making virtue out of necessity. There's less capital. So you've got to be smart about these things. But if we want to create lots of big companies, we have to put more capital to work. We can't let all of them be acquired by large U.S. corporates or even IPO in the U.S. We've got to have money here to invest and to let these companies grow. And so if you look at some of the root cause analysis, I mean, we've talked about pensions before.
8:01In the U.S., the allocation to venture from the pension ecosystem is about 100-fold what it is in Europe. When you look on a percentage basis, there's almost no European pension money in venture. And so that just means there's a massive chunk of capital that you don't have in the system. And it's really interesting when you think about why that might be. And I think we've discussed part of some of the things you could do to solve it previously. But why is the system locked the way it is? And it turns out that if you take the UK, for example, so UK pension funds, they hold more than 25 % of all guilt, all kind of the UK government debt.
8:41And so the government needs the pension funds to keep buying UK debt. And so there's some analysis that says that if the UK pension funds stop buying as many gilts as they do, interest rates would go up. And so there is a conspiracy theory that says that the reason why the pensions are not being set free is because the Treasury actually is quite happy to have these four pensioners being flocked, all these gilts, because we need them to service this dreadful deficit we're lumbered with. And of course, a much better thing to do would have been to put more money into productive assets so we could create a faster growing economy and create more tax receipts and have less of a deficit.
9:19That's a UK thing, right? Rather than a European one? Yeah. So it turns out that some of the rules you have around kind of these, the prudent man rules and some of these systems that prevent more allocation to venture actually are quite similar in the UK, in France and in Germany. and they've all originated from different places, but you have fee caps and you have restrictions around what trustees should do. And specifically to the UK, the rules I know the best, specifically to the UK, when you're a trustee, the prudent thing to do is not to have volatility in the performance of your investment.
9:53Now, you and I, we might say, well, actually the prudent thing to do is to figure out how you can maximize long time returns for your pensioners, but that is not how the rules are interpreted in the UK. then much more narrowly to say you should not have drawdowns and therefore you should have a bigger proportion of your assets in in gilts and bonds and something that isn't as volatile and already there you can start to see part of the issue yeah presumably if you lifted that cap that would that would help with the later stage funding proper right because the pension fund gap is what can fuel a lot of that later stage level of capital i think there i think there's three three aspects to this so one is you've got you absolutely you have to lift that the second thing is what equities do you then invest in if you invest in equities?
10:36Do you invest in U.S. mecca caps if you put money in Europe? And again, it turns out that you get the same tax breaks that are very generous, by the way, more than 40 billion pounds a year is with the treasury spends on pension tax breaks. You get the same tax breaks whether those pension funds are invested in productive assets for the UK economy or whether they're put into U.S. mecca cap stocks. And one could question whether we should subsidize those U.S. mecca caps twice quite as much as we do. That's the one thing. The second thing is, once you've said, okay, more into equities, the question is, which geography do you put the equities into?
11:10And there's kind of the incentive. And the third thing is fee caps, because we know that private equity and venture capital is more expensive to run than listed equities. Now, I'm a huge fan of having more liquid listed equities. I think we should have many more companies go public. I'd love for them to go public in the UK. You're going to have a feeder system. You're going to have growth capital. and if you have fee caps that are very low, you can't afford to put money into venture capital. And again, there's a structural issue with how the pension system is set up. So I think you've got to tackle some of all these things in parallel and then you can start to unlock the capital.
11:43Anything else, Joe, that you'd add? I mean, what other blockers can you perceive? What else would you add to the argument? Yeah, I think the late stage funding is the big one for me. And I think the way pensions are structured here is a big factor for that. The second is the IPO and the exit markets. One factor of that is the later stage funding, because if you have to go to the US to get the later stage funding, then you're going to end up selling or IPOing over there. The other is that, unfortunately, one of the byproducts of the US winning that internet race over the last couple of decades is that the NASDAQ has absolutely ruined.
12:14And there is, again, a small number of very inquisitive companies in the US. I think that's going to take time to unwind. But ensuring we have that late stage funding here, I think, is at least one step towards that. AI is the new internet race, right? Yes. You could argue. So how do we maximize our opportunities with AI? Yeah, and this is what I'm really worried about, right? Because there was a winner of the internet race, and that was the US, and that was great for them. Would it be nice if it could have been Europe? But arguably, it's not the end of the world that it wasn't. There's still huge amounts of value created off the back of the internet here.
12:48The thing I'm worried about with the next phase, which is the AI race, is not that we don't win it. Again, we might do, but it's unlikely. I think the US probably will again. What I'm worried about with the AI race is that we actually lose it by quite some margin. And we're not a sort of second player. That's what I'm most concerned about with this. Some of the other continents are placing strategic bets in AI computes and power. I do think Europe has been a bit too slow and is not acting strategically enough there. When you think about some of the bottlenecks and choke points that are coming, I do think we could be strategically disadvantaged.
13:20And that's a big concern for me. Yeah, it's something we talk about a lot. We do rattle around a lot. I mean, we are going to talk a bit about energy later, which I think will be a massive unlock if Europe can get low-cost energy into business and into data centers, et cetera, et cetera. But the other topic that I want to talk about very quickly now is M &A. H1 has been bumper for startup acquisitions. Why is this important? We're going to come on to in a second, but a few facts and factoids. H125, M &A total nearly 100 billion, up 155 % year on year, Deal count in the mid 400s, which is fairly static.
13:55However, the value is significantly up. So that's over the last three quarters. Deals over a billion and over 5 billion have gone up 17%, 16 and 17 % roughly compared to 24. Obviously, Google buying Whiz was a mega deal. It accounted for a third of that 100 billion at nearly$32 billion. That was the biggest deal in the part. Of the largest 12 deals, two are not US. There's one Oz and one Chinese. So it's not all, but it is pretty much US centric. The other thing that kind of caught my eye was what is an AI business in this context? So 15 million was disclosed as AI M &A, but I don't know how you work out what is an AI company and what isn't.
14:35Mads, start with you. Big bumper H1 for M &A. What is being bought? Who's buying and why? How do you unpack the 100 bill? You've already mentioned some of the big deals. I mean, if we look at three categories that dominate, one is clearly around the code generation. It keeps being the poster child of AI productivity enhancements. And if you look at the new frontier there, well, Windpurf, as you say, was in the headlines, kind of just shy of$3 billion deal that happened with Google buying parts and then, you know, others buying other parts. But the key thing there is Google is behind Anthropic when it comes to market share of LLMs for coding.
15:15So Anthropic is the competitor on the LLM layer. And then you have Cursor that's doing super well on the application layer. And Google is trying to play catch up. And so for them to get some of those windsurf skills, I think were key. Clearly, I want to make a big push there. Google was also very active on the security side with WIS acquisition. Again, you know, when you have more agentic enterprises, you need more zero trust security. and you have Google on one angle finding that AI battle and on the other hand still playing catch up as a hyperscaler behind Microsoft and Amazon. And so they need those security credentials to bolster their fight.
15:56And then the third thing is vertical applications within healthcare, within finance and elsewhere where buyers really want to prove an enterprise traction with AI, so not just tech. And so some of these companies that were bought there already have 100 plus enterprise customers. using AI. So those, I'd say, are some of the three big areas where we've seen the acquisitions. Joe, we have kind of mooted the fact that application layer is where Europe is going to shine. What would you, anything, any of these themes or topics that you would shine a light on or anything that's worth noting or anywhere where Europe can take a play here?
16:33Yeah, I think this is great for Europe, actually. I think the M &A and IPO markets have been cool for a while, mainly due to macro factors, rising interest rates, Ukraine, tariffs, etc. I think that period of wait and see is thankfully coming to a bit of an end now. But I think more importantly, there's been a sort of wait and see period on AI, a lot of hype. But now we're starting to see some things shake out, new categories being formed and some front runners in those categories appearing. Daily, there'll be the headline grabbers, some of the things Mads just talked about. The interesting bit for Europe, given where we are, is actually, I think, that second tier of acquisitions.
17:07Some of the incumbent strategics are now having to place bets to stay relevant, bets in the application layer, and or some of the more exciting businesses on IPO paths need to make bets to form their IPO story. I thought quite an interesting one was the nice acquisition of Cognigy. I think smaller ones would be, for example, Zendesk buying Klaus. I think there's going to be a lot of activity in that second tier of M &A, which is really good for the European companies, actually. I think the other really interesting trend is around AI-enabled services. This is an area we've been investing in a little bit, where I think actually businesses doing traditional services, but with their own AI core, so they can have a step change, better margin and scalability, they're going to start getting bought up by PE, especially as interest rates come down, as roll-up plays for legacy services businesses.
18:00So you can get a margin and a multiple arbitrage play. And again, I think that's probably pretty good for European VC as well. So I think it's pretty exciting, actually, as this window opens up. I think you're going to see a lot of that second tier of strategic M &A. And I think there'll be a lot of opportunity there. Mads, just stepping back, why is M &A important? For those that are, I mean, obviously, if founders are listening, they're kind of deep in it. They don't really kind of pay much attention as to what's going up upstream. But why is this all so important? Everybody's got a boss. Everybody's got a customer.
18:34If you're a founder, you're raising money to build your company. You're raising it from venture capital investors. If you're a venture capital investor, you're raising money from limited partners, which could be endowments or family offices. And even some of those, like fund funds, might be raising money from somewhere else. Everybody's going to get money and capital from someone. And at some point, they're going to expect some money back. And obviously, the time they get that back is when we have some exits. And some of that might be IPOs. Some of that might be trade sales. And clearly, this second piece here, this M &A activity when you have trade sales.
19:07That's when the dollars flow back up, the stack back up to investors so they can deploy new capital to new great things. And so it's a wonderful circle of, if not life, then at least of the venture and startup industry. And so it's absolutely essential. These exits, they're the lifeblood of what we all do. Yeah. And how else would you conclude this? What are the key aspects would you bring out? We've seen some really, really exciting companies come out of Europe and achieve unicorn status here in the last six months. That's not exits yet. That's still paper gains. But tomorrow's exits and dollar gains are created on the back of today's business progress.
19:43And so I think we're definitely seeing that. We've seen N8N. We've seen Lovable. We've seen some amazing companies wave out of the UK. Great, great businesses that will be tomorrow's exits. conundrum that European founders of VCs might face is you know that there's a load of incumbent system of record type players in almost every industry and I'm sure you guys have seen as we have a whole swathe of automation players sitting on top of that creating automations of the workflows and actually now we're getting to a stage where those system of record players will be looking for M &A to buy that layer above to protect themselves and that can be extremely attractive for VCs and founders.
20:23But probably the original vision of that automation there was actually, what's the need for the system of record? We can replace that in another five years, 10 years of growth. And that's the really big win. But how have we got the appetite for that? Because you've got to raise again, you've got to go really into battle with people that have been a bit friendly to this point. And so I think there's going to be some really interesting decisions that boards are going to have to make around, do we take a 500 million billion acquisition from an incumbent or do we actually go again and try and take them on?
20:52I think it's going to be really interesting to see that unfold industry by industry. I think you're connecting the dots here in a beautiful way, Joe, because that is exactly the experience that we have seen firsthand of the difference between a European ecosystem and a US ecosystem. If you want to displace the incumbent systems of record, it is a different level of investment to build that full stack platform from a capital perspective. And it's a different level of investment from a go-to-market perspective. It takes a lot more sales resource to get a Fortune 500 company to rip out their ERP platform than to buy some tools and stick on top.
21:27And so if we want to compete, we need more capital. And I think that's exactly the type of growth capital you were talking about earlier. So I agree with you. We should aim higher. We should go for gold. Well, let's talk about some more money coming through Europe, which is obviously only a good thing. So this week, we've seen Porsche and Deutsche Telekom in talks to anchor a 500 million euro defence fund. This is going to be run by DTCP, a Hamburg-based investment company. Now, interestingly, Porsche has dropped a previous requirement that it couldn't invest in defence. It was only doing dual use by its own terms before, but now it's relaxing its own rules.
22:08Alongside EIF, who are a mega investor across Europe and KFW, they're also relaxing their requirements around defence. Lake Star is also raising a 250 million euro defence fund. And I know of at least three other new and specific defence or dual use European funds that are coming online this year, one in France, one in the Netherlands and one in Germany. So we are really waking up, really waking up quickly. Question for the room, Mads, maybe I can start with you. Is this announcement specifically between Porsche and Deutsche a big deal? And where does it fit in context. I don't think it's a big deal in monetary terms.
22:44I mean, they've got 500 million euros, a quarter of a percent of what Europe needs. But I do think it matters who is doing it and how. I think Porsche and Doja Telecom are total mainstream corporate firms, and that's a super big market shift. If we think back, you know, for 70 years, German industry has been working really hard to avoid anything that looked like or felt like military. And so Porsche is probably backing weapons startups. That cultural shift is probably worth a lot more than the money themselves. We can unpack a little bit what's been going on under the covers. European defense tech investment, I think it was about$5 billion.
23:27So the$500 million is not a tsunami, but it's a sign that the tide is starting to turn. I think that's a good thing. To put that in context of Helsing and Enduro. We've actually seen Enduro founded about four years earlier than Helsing. I think it was 2017 and Helsing in 21. And although Enduro has a high valuation now, I think it's 30 billion versus Helsing's 14, it's taken longer to get there. So Helsing is actually growing a bit faster right now. And that's exciting. But the interesting thing is to get Helsing off the ground, we needed Daniel X money. And we just don't have that many you know founders that have built 100 billion dollar companies around that could fund this out of their pocket money and so that's why it's essential that we get the lease season to get it's all this off the ground so i'd say the money themselves not a huge shift but the fact that it's happening and that the sign that the tide is turning is very meaningful that comes back to your recycling point doesn't it it's not just money that gets recycled it's the actual founders that get recycled as well and then you can do magical things joe what would you pick up on here Would you think this is a good, bad, big thing?
Read the full transcript
24:35I mean, I should start by saying we've not invested in defence tax. It's not an area I know huge amounts about. I think I'd agree with Mads that monetarily speaking, it's not a big deal, this. It's a relatively small amount. But I think it is more interesting as a sign of perhaps there's going to be more of these coming down the line. And actually the signal of intent from some of these big European manufacturers, they're going to start moving back into defence. So I think from a sort of national security, national industry perspective, it's a pretty good sign. You know, defense has always been a hotbed of innovation.
25:06I think that's going to increase again now. And for obvious macro reasons, there's going to be a whole lot of money, government and otherwise, going into this space. So as a financially motivated investor, you've got to look at it. We theoretically do it because we're a generous B2B player, but we haven't backed any defense companies to date and we're not actively developing a thesis. I think that's mainly because selling to governments is difficult. We've always struggled with it. We've always struggled as a generalist with the opacity of military and government customers that you can't speak to.
25:41Because how do you get comfortable with product market fit if you're not an expert? And I don't think those things are going to change. Yes, you can look at dual use, but actually that's pretty hard to pull off, I think, a lot of the time. Because you've got some high quality military grade tech. those customers aren't probably going to want that out in the open domain where people can inspect it and learn about it so i do think generally speaking it feels like an area that will be more specialist but uh yeah that's that's my personal view for now yeah i think i think that's fair i mean the the dual use thing we we talk a lot about defense and then the resilience layer which goes around defense so a lot of the companies that we've invested in over the years mads and i you could easily classify as dual use because there is a layer that could be used in a military application around energy distribution or management of whatever so it almost depends on how apart from kinetics the bombs and missiles and the things that physically kill people where does cyber fit or where do these other kind of more resilience layers and so that's i think there is a a non-ethically challenged layer which i think every country in europe has now been given license to focus on and germany seems to be absolutely crushing it so mads i mean what's going to happen more broadly i think in europe with the european economies and our ecosystem more broadly how do you see this kind of playing out look three scenarios either we'll keep trundling along kind of managed decline low growth not much happening that's the kind of negative scenario that while the optimistic scenario is we'll all start to get a sense of purpose again.
27:18We'll deregulate. We'll invest. We'll focus resources and dollars on the strategic sectors that matter around energy, around AI, around technology. We'll reform planning laws. We'll unlock pension money and growth will go from sort of zero to 1 % back up to, you know, two or three, 4 % per annum. We'll solve the financial problems we have and right off into the sunset. That's maybe a slightly more optimistic scenario. And what will probably happen will be something in between. I don't think we'll manage to go, you know, all the glorious way, at least not in one go. I do think there will be some incremental steps.
27:53I think we've seen Germany say many of the right things. I do think we're still a little bit lumbered with a society that in many ways is quite mature, quite advanced, quite sophisticated. Many people have a lot to lose. There are many vested interests. And, you know, not least we have many pensioners that are very good at voting, of course, and are thinking about society in a certain way. They may be less worried about, you know, what is the growth rate going to look like 30 years from now and more about kind of how do we protect the rights we feel we've built up over our lifetimes right now.
28:26And therefore, we don't want to change too many things and we don't want to deregulate too much and et cetera, et cetera. I do think if we want to compete and if we're serious about regaining the growth trajectory that Europe has had in the past and could have, we will have to make some difficult choices around deregulating, around change, around reform. There will also be some fiscal reprioritization. The tax burden is too high in Europe. Debt levels are too high. So there are lots of things we'll have to look at. If you're asking me what will happen at the office, I can't say for sure. I don't know if anybody can do that.
29:00But you can sort of construct some scenarios and I guess you can put some probabilities on them as well if you're a song client. I don't think it's going to be one. I think we are not going to be in the decline on this level. I think Germany is obviously showcasing a way forwards. The UK has been good on the talking head points. I think there's a lot of goodwill to invest in defence and do more and to think differently about venture. I hope some of that comes to pass. France is obviously also crushing it. Lots of government support, therefore, to unlock private capital. So it feels like it's not going to be one on your list.
29:37It's not going to be like the nothing, but I'd like to see two plus. I dare say it's not going to be this three. Hell, let's go for it. But I'd like to see at least your second option in the plus bucket. But we'll see. Well, listen, we're talking about energy. So let's dig into energy. I was reading, seeing, exploring this week, energy, its creation, its storage, the technology that makes all the magic happen. And it feels like this week there's been a few steps forwards in various guises. I might be having happy ears or happy eyes, but just some background for context. A US facility has achieved a 97 % yield in recycling lithium.
30:22It's been reported that CATL, who are the largest battery manufacturer in the world, they take care of roughly 35, slightly more, 35 % of EV battery market share. They may have produced, and they haven't officially announced this, but they may have produced a$10 per kilowatt hour sodium ion battery, which would be a massive game changer. For context, standard tech is roughly between$75 and$115 per kilowatt hour. If this is true, this is absolutely game changing. Chinese researchers this week have created Supersteel, which will enable the latest advanced fusion reactors. Pulse Clean Energy has a new credit line of£220 million to invest in battery storage.
31:03That's coming online in 2030. Plans have been released for 250 new battery factories across France, Germany, Italy, Norway, Spain and the UK within the next eight years. So they're just some of the things that are going on. Obviously, we've had Tesla looking to power UK homes. We had the BYD Octopus deal only a few weeks ago. So lots of these energy creation, storage, battery tech things happening. So question for the room. Have I got happy years? Are we on the edge of something magical which will unlock magical things for Europe? Mads, kick off. Oh, I love the magical thinking. Look, the truth is, technology innovation continues.
31:42It's a great thing. We've seen that with solar PV. The cost improvement have been dramatic. You know, more than 90 % production since 2010. Battery storage keeps improving. I'm not a great believer that there'll be one silver bullet right around the corner that will fix everything for us. Chinese manufacturing excellence is excellent and they'll keep driving the cost down. And we should just expect that things will keep getting better, harder, faster, stronger. All the things we think are great. And Europe's energy is just way too expensive today. And some of that is due to tech and some of that is due to regulation.
32:16We've seen some initial steps in the UK. there's been a nuclear planning reform, which was announced just six months ago in February. In 2011, restrictions were put on nuclear power plants in the UK, kind of limiting the sites to eight in terms of where a nuclear could be built. They've been removed now. There's been a new nuclear regulatory task force put in place that's reporting to the prime minister. Two and a half billion pounds have been put aside for these small modular reactors. Rolls-Royce is in the picture. I think that's terrific. We've listed a nine-year ban on onshore wind, right?
32:54That happened, I believe it happened last year. And there's been a lot of talk about the great connection reforms. So you can tackle this terrible 15-year wait time we have today. But I do think we need to go further. You just need to be really, really aggressive about ramping up energy because it's so expensive. UK industrial energy is four times more as expensive as US equivalents. let aside China, let's say we can't compete with them, but we are competing with the US, there's no doubt, right? And we can't be four times as expensive on energy. It feels like there are, maybe to your point, there are these lots of kind of little elemental parts that are starting to come together, which when you connect the dots, there might be some absolute movement.
33:36And obviously with AI coming online, so Joe, handing over to you now, obviously AI is a massive energy sucker. How does Europe, how do we think about AI in Europe? How do we square that circle? Yeah, I think to your starting point around happy is, I think I'm a long-term optimist on energy and power. I do believe fusion and things like the small modular reactors will be a big part of the climate crisis solution and an important step to the theory of abundance. And I do still believe in those, but I think they are 10 year plus timeframes still. Still 10 year plus. Yeah, I believe so. And so they're not going to help this.
34:12And we talked about earlier, which I believe is one of the biggest problems facing Europe right now, which is the choking point on AI caused by computes and caused by power. And our power costs and our power availabilities are going to be a real issue if we want to compete on the AI stage. I think the sort of positive news on that is I do think the battery evolution over the last couple of years is really positive. And that's really positive for Europe because we have renewable technologies and we have renewable assets, literally, to solve this power problem. But I think we need to, one, solve some of the red tape issues that prevent the rollout from these technologies.
34:53And two, we need better storage because that's the ultimate limiting factor. So I actually think the news around the sodium battery I found really very encouraging. doing. And I think that could be a bit of a game changer for Europe on that power bottleneck. It doesn't solve the compute problem, but it may at least be helpful in the power bottleneck. The other thing I saw was that massive oil field in Poland that's being discovered, which, again, not great for the long term solution, but could be pretty helpful short term as well. So let's lean into AI corner. Lots of AI news this week. Maz, let's kick off with GPT-5.
35:28What's the good, the bad, the ugly of GPT-5? The industry's been waiting on this model for 18 months. We've talked about it in many episodes. It is an exciting model. It's a unified model that combines reasoning with fast response. And perhaps most excitingly, ChatGPT is the world's main interface with AI today. It's got more than 700 million users. And the new model has been provided free of charge to all of them. Not in unlimited quantities at all levels, but it's effectively a free model. We can go back to some of the monetization and some of the stuff on the back end. But that's exciting, right?
36:04The world is getting better AI. We also know there were lots of controversies around the model. 3 ,000 users positioned to bring back GPT-4.0, the old model, which some people thought were friendlier, and they really got to know it. And they sort of say that the new model is a little bit more like an emo teenager and not quite as expressive. And the initial benchmarks were perhaps a little bit disappointing. It placed fifth behind some other models on SimpleBench. And then there were some model selector problems at launch. There's sort of an auto switcher that sits on the back end and tries to figure out how complex your problem is.
36:43And then depending on whether it's a really simple thing to answer or something that requires more thought, it'll try and route your query to the right model. Well, it turns out that that model was broken on launch day. And so you can imagine when you have 700 million users all trying to get up on this new GPT-5, thinking that they get lots of smarts. And a big proportion of them, for erroneous reasons, get the less smart models. And they think that it's a really smart model and it's not as smart as the past one was. Well, there will be some recriminations. But look, the coding benchmarks are looking very good.
37:17In fact, it looks like it's placing slightly ahead of Anthropix Opus 4.1. Anthropic is the runaway winner in coding when it comes to API use, is the model that sits behind Cursor and lots of other of the pop-in coding tools. And it's also by far the most expensive model out there. And GPT-5 is a lot cheaper. I think it's not quite, it's like 20 % cost of Anthropic's most expensive model. So there is a threat potentially to Dario's business there and to Anthropic's business if GQT5 performs as well in production as it does in some of the tests. So look, it's a big-ish deal because of the size and scale of OpenAI and who they are.
38:01But I'd say, you know, people are also saying that the gains are perhaps slightly more incremental than some of the site would have let folks to believe. Yeah, it feels, I mean, the two things that I like the auto switcher because I was not a massively deep user, but I can imagine that being irritating for many and the speed just the speed of response i think has been has been massively up so i kind of like it i kind of like it as you joe are you a chat gpt user i am and i think from my perspectives with these things i agree with matt it's fairly incremental and it sort of seems whenever one is one of these releases things rejigger it um and depending on who you speak to and what use case they're ones that are better than others i think are we in peak do you reckon no no no So, you know, I'm looking forward to the sort of when we start to move into operational browsers and that kind of thing, you know, GPT operator.
38:56I think there's going to be a whole new wave of usability things when that comes through. Yeah, yeah. No, there are. And obviously, obviously the application piece, which we bang on a whole bunch about, I think is absolutely going to be exciting. What else is on in AI corner, Mads? You've got perplexity. You know, Google has been threatened by the Department of Justice with a partial breakup here or forced sale of the Chrome browser. Chrome browser being very important to their, you know, the Google search business because it informs all their advertising sales and so tells them a lot about what users look at and look for.
39:32And so that's really important information. It's interesting, you know, Amazon has had a lot of stick for being behind in the AI race. And Microsoft, Lava Mohaden, they made an early deal with Sam Altman and they've really been powering ahead. Their cloud business has been performing super well. And although Microsoft doesn't have their own model, obviously they've had that OpenAI partnership, which stood them well, at least initially. We'll see how that goes. People are saying that Amazon doesn't have quite the same thing, but it turns out they're a major investor in Anthropic. They're also a major investor in Perplexity.
40:08So if perplexity manages to buy Google Chrome, that means that through the back door, Amazon, who's got a very rapidly growing ad business online, gets one of Google's crown jewels and will have anthrox and will have perplexity. And I do think business will have a good laugh if that ends up happening. So we'll see. I think we'll still see lots of twists and turns here, but it's an interesting development. I thought this was a bit of a PR stunt. Joe, what do you reckon? Yeah, I mean, you know, Google Chrome has the users, has the cache, and actually has a lot of the data to do the AI browser really, really well.
40:48Perplexity has had a leap on the technology, but Google will close that gap. And so if anything, purely commercially, it would make more sense for Chrome to be buying Perplexity. I think the commercial reason this might have come up is exactly because of the regulatory angle. And if we do believe Chrome is going to get diversified, then it makes sense for perplexity to be trying to make a move on that. More likely, I think it's a PR stunt because they're behind on users and it's a way to try and get a much needed jump. I don't know how much you guys have looked into it. Literally the unit economics.
41:18Some of them, it seems like you're basically getting a discount on the underlying compute, whether it's Microsoft or Amazon doing it. And it's difficult to see how the gross margin stacks up. I didn't even worry about the negative. I mean, you look at OpenAI, but they lost, what,$5 billion in the last accounting year? Pocket change. Yeah, pocket change. But at some point, that's going to have to change. Think about it in this context. So they're right now trying to do a round of secondary share sales, so employee shares, at a$500 billion valuation. And I might have my history wrong here, but I do think the first time in history that a loss-making company has been able to put a price tag of$500 billion on itself.
41:58Good. Do what you can. make hay while it's sunny, baby. I guess it's people betting on the penetration pricing play there. Absolutely. And price will keep coming down. Of course it will. And of course, if they can keep growing top line the way they are, I mean, you look at Anthropics top line growth this year. At the start of the year at one, three months later, it was two. Two months later, it was three. One month later, it was four. Now they have five. They've grown from one to five billion of ARR in seven months. It's nuts. Yeah, it's incredible. Incredible. It is nuts. And also, if you think about, what is it, one in 10 people or the 700 million accounts.
42:30That's a hell of a lot of accounts on ChatGPT. The other thing that we were talking about, Mads, off the call was the 15 % tariff that's being kicked back to the US government by NVIDIA to sell chips into China. And then you were talking about Chips Act 2.0 in Europe and how these two stories could connect and how we could look at, you know, maybe love love him hate him trump might be doing something quite smart here so where would you how would you connect these stories and where would you where would you take this maybe just refreshing sort of the kind of the u.s china vibrate here around the chip so nvidia produces the most sophisticated chips for both training and inference their latest generation of black world chips are completely unsurpassed they're incredible and they're the ones that everybody wants to get their hands on.
43:26Their entire manufacturing capacity for 2025 has already been volatile. They can't make them fast enough. Pretty much, you know, I think both the Biden and the current Trump administration agreed that China should not be able to buy those chips. But what the Biden administration did was they said to NVIDIA, listen, you can produce a cut-down version of a previous generation of the, you know, your AI chips, kind of this age 20, that you can sell to China. It's got much less processing capacity. It's got less memory, but the Chinese should be able to sell that. A ban was then later put on it and Trump is now looking to reverse that ban to open up the sale of these chips to China again.
44:06Meanwhile, the Chinese government, for what they're worth, they're saying, listen, we'd really rather build our own homegrown AI chips industry. They've got Huawei. Huawei, they have their Ascend chips. And so the Chinese government has said, listen, all you Chinese AI labs, Don't buy the NVIDIA chips. You should buy Huawei's chips. Some people have done that. They've tried. And the challenge is, and we've just seen that yesterday, DeepSea came out and said that the release of their new R2 model has been delayed because Huawei's chips were not as good as people thought. Back to NVIDIA. They obviously produced the best stuff.
44:42And Trump has said, listen, you can sell them to the Chinese. And I'm just going to put a little bit of tax on top. And is that a good idea or a bad idea? well, now we're over to Europe and they do how this connects to us. We had the initial CHIPS Act that was pushed through in response to the shortage of semiconductors for the automotive industry. Back in 2021, people will remember there was a huge shortage of semiconductors, meaning that the German automakers couldn't make the cost they wanted. And that meant that the assembly lines had to stop. And it was incredibly expensive just because you couldn't get a hold of those chips.
45:18It turns out that the chips you need to put in those cars are really simple chips. The stuff that goes into a car is tens of bucks per chip. It's really, really simple. It's 10-year-old technology, but you're just going to have it. If you don't have it, you can't make the car. So what happened with the Chips Act was Europe was trying to do two things. It was trying to secure supply of those last-generation automotive-type chips so that the car production supply chain could keep running whatever happened. and they were trying to get hold of some of the new two and three nanometer technology so we could also produce the next generation AI chips.
45:54And TSMC today is absolutely the leader in manufacturing these chips. And they said, listen, we can help you make chips, but it's not going to be the good stuff. It's not going to be the AI chips. We can help you make the old stuff that you need for your cars. We'll make a plant in Germany to do that. And that's gone ahead and that's being built and that'll happen. The European Union then partnered with Intel for the sophisticated stuff. And Intel, obviously, is in a bad place. And that project has now been canceled. And that means that the output of the chip sector, as we have it today, is that we have graciously been granted know-how from TSMC on how to produce chips that are roughly 12 years out of date.
46:37Whereas we don't have anybody in Europe that can produce the most sophisticated stuff. And, you know, where I thought one could connect the dots maybe a little bit is, you know, if you look at Trump's policy, you know, like him or not, but he's been quite good at making deals with some of these different actors and now including NVIDIA that have said, of course, we'll pay the tariff to get access to the Chinese market. Well, we've got assets in Europe. We're procuring a lot of the chips. The manufacturing equipment that goes into making it is made by ASML. Why not create an agreement, an alignment, a deal with TSMC and NVIDIA that aligns around working with them to produce and manufacture some of the chips so you can get some knowledge transfer into Europe?
47:20They're the ones that can do it. And I'm sure there's an economic structure that could be put up around it, whether it's the spurn of a tariff or it's an incentive or something else that could make it worthwhile for all parties. So that was how I was going to connect those different dots. So be more Trump. Is that what you're saying, Mads? Well, a lot of that TSMC stuff that's happening in Arizona actually started on the Biden. And there can be many ways to skin a cat, as you know. He won't remember. That's true. Make stuff happen, right? Don't just sit around and pontificate. We can't wait. That's why we need the next Chips app.
47:52We need Chips 2.0. That's what people have been saying for six months. don't forget well it's in the press people are talking about it but where's the legislation where is the actual initiative and and this is sort of you know where we're saying look we've been talking about this for as long as trump has been in office and you know you don't necessarily need to agree with the man and all his approaches and all his policies but there's action we need action we don't have time to waste let's do it baby well um we don't have a dean of the week this week, but I have been looking at the Italian startup ecosystem and their experience, a beautiful surge in activities.
48:32I want to give a big shout out to Italy. New data showing that there's been a record-breaking H1 for investment in Italy. So far this year, startups have raised 655 million euros, beating the full year of 623 million euros back in crazy, crazy 2020. So well done, the Italians. Not only is it a big rah-rah for the Germans, but a big rah-rah for the Italians. And I cannot wait to be able to say a big rah-rah to the UK, which is coming. Well, Joe, thank you so much for joining us. You've been awesome. Mads, enjoy Greece. Both of you, catch you very soon. This has been Upside. See you on the flip side.
49:10It's done. Cheers, Mads.
49:20Upside!
From the publisher
Welcome back to another episode of Upside at the EUVC Podcast, where Dan Bowyer, Mads Jensen of SuperSeed and this week’s special guest Joe Knowles from Smedvig Ventures unpack what’s happening in European venture capital.
This week: Are we doomed to lag the US in wealth creation or is Europe finally closing the gap? Why a record $100B of M&A matters for exits and recycling capital, and how founders should think about selling vs going for gold. Plus: Porsche & Deutsche Telekom anchoring a €500M defence fund as Germany drops its taboos, the scramble for cheap energy and battery breakthroughs, and what GPT-5, Perplexity, and Nvidia tariffs tell us about Europe’s place in the AI race.
Here’s what’s covered:
00:48 US vs Europe in Wealth Creation: Compounders, unicorns, and Europe’s capital efficiency.
09:07 Late-Stage Funding Gap: Why pensions and IPO markets hold Europe back.
17:36 M&A is Back: Google’s $32B Wiz deal, Windsurf drama, and Europe’s “second tier” opportunity.
25:52 Why Exits Matter: Recycling capital and the venture flywheel.
27:09 Defence Tech Goes Mainstream: Porsche, DT, EIF and Germany’s cultural shift.
36:18 The Ethics Question: Dual use, deterrence, and uncomfortable truths.
41:22 Energy Corner: Lithium recycling, sodium-ion batteries, and Europe’s 4x US energy costs.
46:44 AI Needs Power: Grid bottlenecks, red tape, and planning reform urgency.
51:10 GPT-5 Launch: Unified model, user backlash, and coding benchmarks.
54:37 Perplexity vs Chrome: PR stunt or regulatory opening?
58:32 Chip Wars: Nvidia tariffs, Huawei delays, and why Europe needs Chips Act 2.0.
1:05:59 Shoutout to Italy: Record €655M H1 startup funding.




