This Week in European Tech: Europe's next trillion-dollar company won't look obvious

0:00Before we start, five weeks from now, Love Tomorrow Summit takes place at Tomorrowland. Take a look at your screen right now, because there is something amazing in which you're having right now. EUVC is on stage, keynotes on Thursday at the Rose Garden stage and the Investor Lounge at the Impact Circle on Friday the 24th. The speaker lineup is nearly locked in. If you're active in impact investing, sustainable capital or climate-focused venture, the people in that room on Friday are exactly the people you want to be meeting. Link is in the show notes. If you're coming, I would love to see you there.

0:30What I find very interesting to observe when there's a 7 billion put into a model company in China and the state actor puts that money in, all of our reaction is, oh, of course, that's how it works. Imagine if the state actor in Europe put 7 billion in Mistral. There'd be uproar. What uproar would there be? Raging in the streets, isn't there? A few fun, but also maybe depressing facts. The EU generates more startups per year than the US and venture returns in Europe run about 6 % higher. But 30 % of our winners head off West. I guess no great surprise. I know someone in the European government responsible for implementing LLMs across all government services.

1:04And what she told me is, yeah, we're not allowed to use U.S. solutions. So we need a European solution and that's it, you know. So to me, Mistral is in a beautiful position that there's a long queue of customers wanting that product. They just need to get it right and, you know, revenues are there for them. What it means for Europe is we have to wake up and accept that if we don't have the capacity to stand our own defeat, we're not going to get a seat at the table. We're in this new era where there is this technically capable population all around us that is going to enable or be part of the acceleration of this kind of war mode.

1:37Okay, all right, we're on our own. We're going to build, we're going to do, we're going to integrate, we're going to make this happen. Probably the next trillion dollar European companies are already operational and running, and they're growing. They're there. And I wish I knew which ones they were because I would put all my money on. Yeah, let's go for it. Hello and welcome to Upside, where every week we dig behind the headlines that are going to affect European venture. Today, it is Mads, myself, and we have a very special guest who I'm going to introduce any second. And what are we looking at?

2:03We've got pulling Anthropik's fable, technical or political. Some AI chat behind closed doors at the G7, digging into that. DeepSeek raising 7 billion. What will this mean for Europe? Can Europe build a trillion dollar company? Predictions, deals of the week and what's going on in the week ahead?

2:34This show is not investment advice and the hosts of this episode may be invested in the funds and companies featured. Now, today we've got a very special guest on who is one of the people who decides which European venture firms actually get to exist. Joe Schorge is the founder and managing partner of Isomer Capital, one of Europe's largest funder funds backing the VCs that back the founders. Six vehicles spanning fund investments, co-invest, secondaries, writing checks from 1 to 15 million, with 59 unicorns across more than 3 ,400 portfolio companies, including the Who's Who. We've got Eleven Labs, ICI, Lovable Aura, Poolside, Fractile.

3:14I could go on. Joe, how the heck are you, sir? I'm fantastic after that intro. Tell you what. Thank you so much. I'm a big fan of the show and super happy to join you this week. Oh, thank you so much for joining. Mads, how are you, my good man? Great. Summer is here. What a terrific week. I can see you got some sports news in the lineup. I can't wait to talk sports. I had to stick it straight up front because I'm not a footballer. I know you're a football dude. Joe, are you a football fan? Not really. No, I'm with me. I always feel like I lack the manliness when I can't talk about the football stuff.

3:50However, when it comes to England, obviously England winning 4-2 against Croatia, I can't not be proud. But Mads, you were watching it live, right? I mean, I saw you chatting in Slack on the channel. So you were excited. It was just a big thing, I think. You know, best performance from the team in, I don't know, 25 years. And it kind of feels like finally the team has the courage to play their way. It's coming home. It's coming home. This time. Even I remember that song. I think it was 96, that song. I've got some quick news up front and I wanted to basically look at some economic health indicators, some over the pond and some here.

4:28Just to set the scene, please chime in at any time, gents, if you have any thoughts or opinions. But we know that we are in kind of record market territory, but there is a whole world of other things going on underneath that. But the S &P had nine straight weekly gains up until the beginning of June. Now, it softened slightly into the seven and a half bracket, but up about 24 % over the last 12 months. Kevin Walsh, the new Fed chair, held rates at 3.5 to 7.5. However, the consensus is that there will likely be a rate hike or two later this year, with most economists, 75 % at last count, now warning a 20 % S &P drop is on the table.

5:09So we're back in that kind of AI bubble turf. I don't before we go there. UK FTSE closed again, still record highs. 17th of June, 10 and a half, up 19 % over the last 12 months. GDP slightly down, but Q1 actually beat the US, Germany and a shrinking Eurozone. IMF predicting a 1 % growth of 26. Not great, but not recession territory, we could argue. Inflation 2.8, but heading up because of the energy crisis. Rates over here held at 3.75. However, as we might well see, possible rate hikes. So just for broad context, US resilience, stronger growth, but higher inflation, stalling Eurozone, contracting in Q1, inflation fairly steady at 3.2, and the UK somewhere in the middle.

5:55So are we entering the energy shock in better shape than our neighbours? I would say yes. But any thoughts before we move on to the main docket? Mads, anything from you on the broad macro? I think we have a new prime minister this morning, don't we? What? Brennan, he won. Did he? Is this, I didn't see the news. I saw that he was likely to win Maker FILED. He did? More than 50 % of that, yeah. And I'm sure, Joe, this means nothing to you. Do you know about the Andy Burnham leadership challenge stuff? Are you tuned into that kind of side of the UK news? Of course, of course. I heard it a bit on Radio 4 this morning.

6:34And I mean, they're still deciding if they're going to have a leadership contest or not, right? I don't know. It seems to me the UK spends a lot of time changing its government and less time moving forward. Yes. It's all a bit naval gazey. And I saw streeting was up for a fight as well. And who knows, maybe rain will come out of retirement. I mean, this is just an absolute mess. I don't know how we tie that back to European venture, but I think what we should probably do is, Mads, you snuck something in the docket this morning that I thought, and I was challenging you on not the relevance, but the underbelly of this story.

7:12And it's basically Accenture's earnings out. Unpack this. Tee this up and then unpack this. So many people view Accenture as kind of a classical, old school outsourcing company, old economy. This is your dad's services business. Who cares, as some people would say. And still, we teed it up last week as something to watch out for this week. And the reason is that Accenture, in many ways, is a bellwether for what's happening. in the IT services industry as it is affected by AI. And why is that important? It's because enterprise is important for AI and Accenture is enterprise and enterprise technology and enterprise deployments.

7:52So the thesis here, Mads, just so that I'm right. So the thesis here is, as we now move into this kind of application layer of AI, we need to pay attention to these bellwriters. Is that the thinking? Well, 100%, because Accenture has always been about the applications and about implementing the applications and the big transformation programs for large enterprise to make technology valuable. Now, if you go back 15 years, people were saying cloud, cloud is going to be the end of groups like Accenture. Because it's going to be so much easier to deploy stuff and companies are just going to do it themselves.

8:20And Accenture said, well, guess what? And then tripled their revenue over the coming decade. Now, Julie Sweet, who is the CEO of the company, again has said, well, AI is going to be the next tailwind for us. But what happened in the results yesterday is that the numbers disagreed. Revenue was up a little bit, but earnings couldn't meet the expectations. They were mainly driven by cost cutting. Again, earnings per share beat a little bit, but it was on cost optimization. And bookings, so the new contracts that are signed, the forward-looking piece, were down and missed. And so markets, they sort of said, thank you for that update and continued to pummel the shares, which was down nearly 20 % after the earnings were released over the last 24 hours.

9:07Is that inertia, do you think? Is there an aspect of inertia baked into this? Will this bounce? I guess we'll see over time. But what do you think? Well, that's the whole discussion. Is this a transitional piece or is it secular? And in many ways, you know, there's this big discussion about, you can say that the three main components of a large company like Accenture. It's a big business. They do$70 billion of revenue annually. We like talking about market cap. This is real revenue, right? It's a real company. Something like three, 400 ,000 employees. It's a massive behemoth. So the question is what happens to the different components of their business, which is they've got kind of the high margin strategic consulting, the C-level relationships, the one throat to choke, all the stuff that the rainmakers that are driving some of the kind of the upfront piece of advisory work that then can lead to downstream two things.

9:57The first thing is implementation and transition work. And the second thing is sort of ongoing outsourcing and maintenance. What we can see is that the outsourcing and maintenance piece is being threatened by AI because it's becoming easier and cheaper to maintain things with AI. You can do more with fewer employees. And so that's hitting sort of the body sharp aspect of the pie. But the other piece is the implementation. What happens there? And historically, application providers like SAP have relied on large systems integrators for implementation. because if you are an application developer, most firms have said, look, we don't want to have tens of thousands of consultants on our books to implement our software.

10:39It's low margin stuff, let somebody else do it. And it essentially turned out to be really, really good at that. But in the new world with AI, you can implement software in many ways more easily because you don't need maybe 500 or 1 ,000 people for large implementation. You can maybe do it with 50 or 100. And that feels much more like creating IP, like proper software, and you can run those businesses differently. And that's why the likes of Palantir are really successful right now. That's why we are seeing a lot of work from, you know, not just open AI and Anthropik, but also Mistral here in Europe, implementing and working directly with corporates to implement the new AI technology.

11:18Now, these companies are taking the cream today. They're not necessarily going to take the long tail, but it is going to pressure the bookings and the growth of large systems integrators like Accenture. So we're seeing probably a secular shift of value from the companies that do the implementation to those that have the IP. And that speaks well to the likes of the Mistral, which actually, maybe, Joe, you'll have a view of that. I think that's one of the ones in your books, isn't it? We're not in Mistral, unfortunately. I hate to say it. It's one of the few big things in Europe we didn't manage to catch.

11:52Would you go with the secondaries fund there, right? Will that maybe perhaps sneak in at some point? We're always hovering around the edge to buy a bit of that. Yeah, that's what secondaries is great for, right? You can buy into companies that are very successful later. But can I come back to your points, Mads, which is I wouldn't count out yet the implementation phase. I think just because we have a great new technology doesn't mean that every large company in the world rolls that out quickly and easily. And that's what the accentures of the world are there for. And I think that wave is yet to come.

12:28We're still in the moment for, don't forget, you're on the pointy end of the stick. And most of these big corporates are years behind, and they will need all these implementation services, even if it's just at the beginning, the strategic part, what do we do? The other bit, I think I would add to your two pillars that there's a third pillar, which is the strategy M &A part. So if you're a large corporate and you think that your product's being disrupted by new AI competitors, you're going to do a big strategy piece and look at buyer build decisions. right so maybe we should not try to build a new product and wait for our lunch to be eaten we should just go out and buy and i think that's also where the strategy consultants make a lot of money and pick the targets do the m &a process and so on so i i wouldn't count about yet i having having no no a hundred percent i don't think it's a melting ice cube i don't think they're going to disappear completely but i do think it is a threat to sort of continued growth i mean you've seen over the last decade, you've seen kind of a 3x revenue growth, right?

13:29Very solid. For a services company, after all its bodies, you need to hire all these people and they build this massive organization. I do think that will shift a little bit and I don't think they'll be able to grow quite as fast going forward the next 10 years as they have the last 10, 15 years. I want to move into the models because I guess there'll be nobody listening to this show that hasn't seen that Anthropics models were pulled. Three days after launch, the US ordered Anthropics to switch off fable, vibe and mythos worldwide on national security grounds after a jailbreak enabling serious cyber capability.

14:03The order targeted foreign nationals, but the model went dark for everybody. Anthropic is still fighting. They were in the room with Trump and the leadership team, which we'll talk about later. But as we know, the Trump administration is not a massive fan of Dario and Anthropic. But Mads, if you can tee this one up, I personally looking at this, I was not quite sure whether this was a technical thing or a political thing, looking at the capabilities coming through from OpenAI and other frontiers. So is it technical, political, both? How do you think about this one? I think there is a little bit of almost historical reaction from everybody.

14:44And to some extent, you could say Dario's own fault because he's going to been doom mongering and sort of fear mongering about jobs. That's his PR, isn't it? That's his shtick. every time he opens his mouth. He is the one who has been saying AI is dangerous. We need to control it. And guess what? Now he's the one who is being controlled. And he was sort of, he was complaining in retrospect that, you know, the guys didn't do that to GPT 5.5 and OpenAI also has a good model. And why isn't that being sort of controlled the same way? So there is sort of peak hysteria right now, I think, from everybody.

15:17I mean, everybody's pointing fingers at everybody. But when you go under the covers, you know i the taxes will come again and i think they will they will sort this out i think the real lesson is maybe more boring which is don't rely on a single model and you know people talk about sovereignty and sovereignty is very part of it but it's also just concentration risk i just don't be reliable relying on just one provider that my assumption is not not much would have happened in whatever three days right i mean nobody's going to be that reliant on something that's been out that short period of time.

15:50Here's a story from one of our portcos that are doing stuff that's right at the edge of AI, super cool. And they're always saying, look, we develop stuff and we develop stuff with the next model in mind, because the models of today can't quite deliver this enterprise functionality that we want to deliver, but we're almost there and our harness is getting us almost there. When Fable was released, they dropped that in and suddenly the whole thing worked. Yeah. Okay. So that was one model that shifted everything for them. And then of course it was yanked, you know, a couple of days later. And kind of what, what situations does that leave them in?

16:28And what they say is, look, we just, you know, we are, we are now retraining our own models. We're taking the open weight models and we're fine tuning those because we don't want to be in a situation where we're dependent on it. So I think what it does, if anything, is it just highlights how important that is. and to some extent weakens anthropics and in many ways the US's position a little bit. So to kind of this trade policy, I don't know whether it's maybe shooting themselves in the foot. Joe, with LP hat on, does this make the whole European sovereign AI bet more investable? Absolutely. Yeah.

17:04I mean, I'm with you on all of this that systems get fragile when you have single points of failure, right? That's just a general thing in software, but also, I mean, we have one straight in the world that affects the whole energy market. That's a very fragile energy supply chain when you think about it. And the globalization over the last decades has led to fairly fragile systems. So I think this trend to what we're all talking about now, tech sovereignty, is really a move back to redundancy built into systems and so on. And that plays right into the hands of your portfolios, our portfolios. If you're sitting in Europe, you're a large corporate, you're a government, whatever, and you want a European solution, you're going to look to the technology companies building those solutions.

17:51So it's in a way, it's an amazing, you might even go so far as to say it's a gift to the technology builders of Europe, right? And we certainly see in our portfolio companies moving faster and the customer base moving faster. So if you're selling into government or maybe the extreme is the defense side, right? Extremely long procurement processes that go for years. And all of a sudden, there's an urgency to that that's turned years into months. And that's happened very, very quickly. And you see it only when you talk to the salespeople within these startups. So, yeah, I believe that this is very positive for European tech companies.

18:35we're seeing the effects of it already and I think it's got got some years to run which means place your chips now you know make your bets in Europe because I think I mean I'm always optimistic about European tech but I think even even more now with a burning platform to to put solutions in place I think you're right I think that there is also an underlying aspect that I've mentioned before on this pod, but there's something very powerful about this kind of war mode where we're being reminded that we need to think about friends shoring, on shoring, reshoring, whatever that might be. But there is also this underlying aspect where many of the people in government, in procurement, in business, learn to type on glass and are technically capable.

19:21It's not the boomer generation that maybe aren't so comfortable with tech. We are all now working with people slightly older, slightly younger than us on this call who get it and can make faster decisions and know how to play the game. So I think we're in this new era where there is this technically capable population all around us that is going to enable or be part of the acceleration of this kind of war mode. okay, all right, we're on our own. We're going to build, we're going to do, we're going to integrate, we're going to make this happen. Well, maybe that's a fundamental that's driving faster adoption, but I would also say that the normal rules apply, right?

20:05Companies have to build really strong solutions. It's not good enough to have a solution. You have to have a really great solution as always. And so the customer reacting is great, but you have to also be there with a great bit of software, great service, whatever. It reminds me, I was last week at Superventure in Berlin, and there was a new day, a new format dedicated day to tech sovereignty. It was a new thing that we tried. Yeah. So it's gotten to be that big a topic that it needs its own day. And what was interesting to me is tech sovereignty is the whole supply chain discussion. It's resilience.

20:46It touches all supply chains, all sectors of activity. But in Europe, because we have this urgent need for defense, if you sat in the audience of that day, actually half, 60 % was all about defense. And so when I got up on my panel, I said, well, defending a thing is great and that's urgent and we're building wonderful solution. And the progress there is amazing, far faster and bigger than I could have thought a couple of years ago. So, you know, there's protecting a thing, but then there's the thing itself. So we didn't spend enough time on that, right? So we need Mistral to be successful. So we have a European, I know someone in the European government responsible for implementing LLMs across all government services.

21:33And what she told me is, yeah, we're not allowed to use U.S. solutions. So we need a European solution and that's it, you know. So to me, Mistral's in a beautiful position that there's a long queue of customers wanting that product. They just need to get it right. And revenues are there for them. I think we'll have in a year or two, we will have very different conversations on this topic. And I look forward to those. Talking about AI and models being switched off and the whole sovereignty piece behind the scenes at the G7 this week. all the big frontier teams and the Trump admin, all the people fighting, the Amadeus, the Ultmans, well, Sabis isn't, isn't he?

22:17He's kind of sidelined to this. But they were all in the same room pitching the idea of a US-led AI coalition to shape the industry rules, standards, etc. The same day, funnily enough, China's foreign minister re-pitched, because he said the same thing about a year ago, re-pitched the same idea, but with a global AI coalition instead open to everybody rather than just the US and friends. So a question to you both. Of course, the US is leading AI. Everybody knows this. But how should Europe think? How should European leaders respond to this? Joe, I don't know if you've got any ideas on this one.

22:55It's just a big topic. Yeah, a big one. It's a big, scary one. It's a big, scary one. I'm not a policymaker. Europe has to get a front seat at the conversation for sure so yeah you know we're we're used to in europe talking ourselves down all the time but this is something we really can't afford to to not be at the front of so i think less these days about u.s versus europe versus china particularly the u.s europe if you look on the ground you see technology and talent and capital yeah it's a continuation isn't it yeah yeah it really isn't us or them. It really is, you know, something starts here and goes there.

23:35Something starts there and goes here. Much faster and easier than it used to. And some of that is rules have changed about that. But also some of that is just how networked we are. If you build an attractive piece of tech, everybody knows about it really fast. Yeah, it's global. Yeah. It's gone global. So I, you know, I worry less about, you know, will those things flow around naturally? But certainly on the policy side, I think we should be part of any coalition that starts. And probably ideally, it involves everyone. The old deal seems to have changed. There was a deal between Europe and the US, which was sort of along the lines of, well, you know, we're happy for the US to be the leader and to be dominant.

24:18And the US is happy to support Europe and give us a seat at the table because we'll give market access to European companies and sort of everybody wins. And that seems to have changed a little bit, especially with the current administration, which has a slightly more sort of might is right attitude and where there's sort of been a view that, you know, you really have to carry your own weight. And so you could argue both sides of that. But I think what it means for Europe is we have to wake up and accept that if we don't have the capacity to stand our own to feet, we're not going to get a seat at the table.

24:52And so, Joe, to your points earlier, we have to build good products. It's not enough that we are just saying, great US companies, we have wonderful markets here, we'll let you sell your stuff. And in return, you're going to listen to what we say. We actually have to compete and be the best at doing stuff. It feels a little bit transparent, this idea of this alliance that's being proposed here by Derek Amodei. And he had Demas Hassabis in the room. I don't know if Demas is really part of this or not, or whether he was just invited along. sideways right he tends to not get involved in this kind of political stuff but i guess he but he feels very strongly about how ai is applied in the real world so maybe that was the that's part of the key i mean so i mean let's let's steal man what people are doing they're sort of saying look you know there are really two different ways of thinking in the world there are some allied democracies and they should have the frontier chips and they should work together to make sure that ai and the best ai is available to democratic countries instead of western countries and that Yeah, it seems possible.

25:56It makes sense. If you look at the steel man on the other side of why this may not be such a good idea is that China is really leading in open weight and open weight models, and you can't really embargo them. And the risk is if you try to embargo, if you try to contain China in this way, they're going to build their own ecosystem. They already are dominant in open source, and it just creates an even more formidable competitor to the US, so the US ends up losing. And then you may fragment the kind of the tech stacks underneath. If you were on Broadpick and looking at this, you're probably thinking, look, it's great if I can have fewer rivals, that's good for me.