E704 | This Week in European Tech with Dan Bowyer, Mads Jensen & Lomax Ward

14:22Mads Jensen:So I feel like that is an opportunity for startups. A little bit of cold water, but perspective is that the U.S. is also doing this in a big, big way. right and dural now the kind of their own lighthouse new prime new tech defense tech company is in talks to raise a six billion dollar round at the moment they've raised a two and a half billion dollar round i think last year so the numbers again it's like reminds me of open ai and and the other lands like the numbers are very very different europe us but if you just zoom in on Europe. There are opportunities here. Now, I think finally, as a venture investor, I often think about this because I have made investments in defense and I have made investments in dual use is one of the worst things you can do in venture is start writing checks into things that I call current soup du jour.

15:17Mads Jensen:It's like investing in health during COVID. It's investing in the current thing. You want to be investing three, four years ahead. And one thing I'm still trying I figure out is in defense, are you at the risk of the same thing? But I sometimes think actually not, because actually there's so long to go in terms of the integration of these new technologies, in terms of the re-industrialization of the European military industrial base complex, that actually there is a lot of opportunity there.

15:46Dan Bowyer:No, I mean, thanks for opening the topic, Dan. And I think Lomax talks a lot about sort of the details and the implications, but I do think it it's worth just stopping up and considering how different of a world we are in today. If you go back four years, I mean, we had 2022, Europe, our security was guaranteed by the US. Our automotive industry was seen as the world leader. We had the cheap Russian gas. We were completely reliant on US tech and sort of thought that that's going to remain that way forever. All those things are just gone now. And at the same time, we have China that's been catching up in an incredible way, both on the automotive front, but also on the AI front, which is kind of the automotive side is terrifying the European industrial base.

16:35Dan Bowyer:The AI front is terrifying the Americans. And so the Americans are pivoting out of Europe and sort of, you know, there's this US-Chinese belligerence. And we're still trying to figure out what our response should be in Europe. And I think there's a risk that we are overrun on the manufacturing side and the AI side simultaneously. So the stakes are really existential, and that means we can't have an incremental response. So I think that's the backdrop. The upside, if you will, is that we now have a license to put sort of those luxury beliefs we had to one side. We're going to go back to that later.

17:10Dan Bowyer:I think we're going to talk about some of the AI companies and how they are changing their views in one of the next segments. But I think for us, it feels now we now have a license to focus on competition, on actually being the best at something and not on getting distracted by these other extraneous activities that aren't really creating value, but we're just sort of navel-gazing and doing other stuff. So I think that is possibly something positive that could come out of this. in terms of what's ahead for us. So if you think about Ukraine, we're four years in and Dan, you talked about the incredible human toll.

17:45Dan Bowyer:I mean, I think it's 1.8 million people sort of combined casualties on both sides. And it's just, there's not been this level of war that we've seen since the second world war, I think. And it's kind of a staggering number of people dying. And that's a big tragedy. And the question is, how do we resolve it? And I think we owe a lot to the Ukrainians for standing up to Russia. And the least we can do is to support them economically. And the Trump administration has obviously said, you know, the US, we're out. And so as Europeans, we're the ones that are left to step up and step in. You know, from what I can tell, Ukraine needs about 60 billion euros or dollars a year from partners to sustain the war effort.

18:31Dan Bowyer:And Europe has been stepping up, but still only about 30 billion a year now. So there is a gap that needs to be closed to be able to sustain Ukraine. Now, the Russian economy is struggling. Growth has certainly collapsed completely, and it looks like they're going to go into the reverse. They've been spending down on their sovereign wealth fund. The oil price for Russian oil is very low. So there are real challenges ahead, but it doesn't look like this is going to resolve quickly. And so there is a real question of who's got the will to sustain this for longer. And I think it's very, very important that we stay in this and we continue to support Ukraine.

19:11Lomax Ward:I was reading this week how the Ukrainians are measuring the financial impact they're having on Russia. And I can't remember the name of the metric, but they basically measure they have like drones that whatever, destroy shipping. and so far their run rate is for every dollar they spend on whatever piece of kit, they're destroying$100 worth of Russian equipment. So I think obviously while they're operating at that 100x, I think hopefully there will be a point where the Russian war machine slowly runs out of steam, but I don't know. I have no context for how long that's going to take, but I do hope it's a relatively speedy close.

19:57Mads Jensen:Well, wars are basically economics, right? Yeah, yeah, yeah. No, I would just close. I would close like Mads rightly focuses on the terrible tragedy that's going on in Ukraine. I would say if you want to try and take some positives away from this going forwards. So firstly, as Europe has had the wake-up call that it needed probably in this regard, and is starting to spend. You look at the European rearm program by 2030, they want to spend 800 billion over 10 years. So that's real money. Secondly is what's gone on in Ukraine is the Ukrainians are the most up-to-date cutting edge. They are fighting a war of the 21st century.

20:43Mads Jensen:And so actually that skillset knowledge resides within Europe now and actually not the US, for example. So there is, you know, I know Ukraine is not in the EU. It's partly why they're fighting this war. Well, it's trying to be. It's partly why they're fighting this war. It keeps asking, doesn't it? Yeah. But let's just call, you know, Ukraine, Europe. And it is very integrated, obviously, in a military level with the rest of Europe and NATO. So I think that, I mean, I wouldn't call it exciting because it's probably the wrong word to use, but it's certainly compelling.

21:15Lomax Ward:Yeah, there will be a legacy. there will be innovation manufacturing startups from our investments in in defense is they're

21:25Mads Jensen:deeply integrated with these absolutely crack us ukrainian teams right yeah you know so yeah

21:32Lomax Ward:i was in i was in i was in a drone factory last year looking at some of the tech they were building and and at that point it was they were buying in all kinds of weird and wonderful stuff from china from different weird and wonderful tech pieces of kit to construct these drones. It was quite an eye-opener. On that note, actually, there are two articles worth reading.

21:53Mads Jensen:One is a great article in The Guardian last weekend about the origins of the war and how the UK and the US saw it coming and no one would listen. Not even the Ukrainians. Really, really good long read. And there's a good article, actually, in the FT today about exactly what you're saying, Dan, which is around... Because one of the problems with defence is also just scaling up these supply chains. and actually manufacturing at scale and at speed. And actually in the context of modern warfare and drones, actually this stuff becomes obsolete incredibly, incredibly quickly. So, you know, at one point...

22:23Lomax Ward:The rate of innovation is nuts.

22:24Mads Jensen:So at one point, at one stage, you're like, oh, we need to make loads of these things. But at the same time, it's like, really? But then we'll have a stockpile of stuff that's out of date. This is not like... Yeah, yeah. And also the Russians work out of counter, don't they?

22:36Lomax Ward:So they're using all the fiber optics. They're running 20, 30K fiber optics to run these drones. I was also reading about... I can't remember the name of the thing. the they're like 2000 kilo ships the the water drones they used to go and basically run at run at the the russian fleets to sink the russian ships and how they're how they've managed to work at how to get short-range missiles to fire from these tiny tiny vessels while they're rocking around and all the salt water going into all the electrics it's like it's quite amazing stuff although unbelievably horrific in the same breath right let's talk about there is a bunch of news this week about obviously AI is always going to be in our docket.

23:15Lomax Ward:It's super exciting. But this week has kind of been safety disruption, some naughty distillation from the Chinese open source models. Starting with safety, Anthropic was founded specifically to be the safe alternative to open AI. This week, it updated its responsible scaling policy to say it will no longer delay AI development if it lacks a significant lead over competitors, which basically means safety is now unconditional your xai has agreed to an all lawful purposes standard which is a nice you know broad blanket statement grok is now cleared for classified intelligence which scares the living bejesus out of me uh weapons development and battlefield operations the pentagon has threatened to brand anthropic a supply chain risk if it won't lift all safeguards so a lot is happening on this kind of safety piece, Mads.

24:06Lomax Ward:Can you kick off this little bit?

24:10Dan Bowyer:Well, I think it ties in very well to what we just talked about, which is that existential threats change context and change beliefs. Now, we'll remember Google in 2018, they had this discussion about Project Maven, which was a defense and intelligence project from the US government at the time where Google employees pressured the company to basically give that up and say, well, you know, we're not going to do it. We're not going to support these things. Now, four years later, ChatGPT appears. Suddenly, people thought Google might disappear overnight and OpenAI might take their business. And then today, Google has major Department of Defense contracts and kind of the existential threats completely rewire that moral calculus.

24:53Dan Bowyer:And it's sort of, I think, a little bit of an analogy to what's happened in Europe, right? It was a time going back 10 years when we thought, listen, we are so rich and so wealthy and all the rest of the world, the Americans are also wealthy, but the rest of the world is just developing economies and they'll never catch up. And we can sit here and just sort of think about our regulation and how to be good human beings and not worry too much about competing too hard because nobody can catch us. And suddenly things become a lot more existential. And we're reminded that actually, you know, there are some of these survival things that that do need to go to the top of the priority stack because if we forget that, well, it's going to be a whole different conversation we need to have.

25:33Dan Bowyer:And now this week, the same has happened to Anthropic. And that's no different from what happened to OpenAI previously. OpenAI, they opened out as wanting to be in an open company and develop AI to the benefit of humanity. And suddenly they worked out that they were probably on the trail of a trillion dollar valuation in a massive company. And all the stakes changed and the calculus changed. And Anthropic is now probably in the same place. They started as a safety first company, but suddenly they have these pressures from all these sides. On one hand, you've got the Pentagon, the Ultimatum. You have Chinese open source models.

26:09Dan Bowyer:The Chinese is possibly distilling on Anthropics models. And you've got all these new deep sea model that's coming out this month. Sounds like it possibly has been trained on black wells that have been smuggled in. Right. So on one hand, you start out having all these good intentions, but then reality catches up with you. And, you know, some people will say it's a tragedy. I think it's just it's human nature when the context changes and we're under these pressures, then the kind of the value of the moral compass can change and evolve. And I think that's sort of that's sort of possibly what's happening here.

26:46Lomax Ward:It's kind of sad, though, isn't it, really, that we're going to end up with this? I find it sad. I know that we're trying to be pragmatic as well as think from the human angles. but it just feels like we're just taking 10 steps backwards with all this nonsense. But Lomax, you've got some notes on this point as well, haven't you?

27:01Mads Jensen:Well, no, I'm just like, it's probably no coincidence that probably only a few weeks after you've announced the new$30 billion fundraising round that you're now stripping back the guardrails from your AI, right? I think it's, you know, call me cynical. And then in regards to the specific point around Anthropic and the DoD, which was that Anthropic was selling it out of the only LLM model company before Grok, as you mentioned, with a contract into the US DoD for 200 million a year, I think it is. And Hegseth has basically said they had two restrictions. One is that the models could not be used in the development of autonomous systems used for offensive purposes, basically without a human in the loop.

27:45Mads Jensen:So drones or autonomous systems that would kill people automatically or for surveillance of the local population or the US population. And they were kind of restrictions on the use of the anthropic model. And Hegseth basically said, got very aggressive and said, if you don't restrict though, if you don't take those out by, I think it's Friday this week, you guys are out. In fact, not only are you out, but you're deemed to be the same level as any kind of Chinese corporate that could be in a supply chain. So very aggressive stance. So for me, that one, I'm like, well, to be honest, Anthropic, if you want to sell software to the USDOD, that's like, you can't kind of have your cake and eat it.

28:25Mads Jensen:That's a very, very clear one to me. It's like you can't sell into the USDOD, especially at the moment, what's going on in the White House? Like, what did you expect? And then again, you raised 30 billion. I don't think the investors in that round are going to be happy for you operating with a bit of a handbrake on when your competitors aren't. So I'm just not surprised. It'll be worse than that, Max, won't it?

28:47Lomax Ward:It'll be worse. I know you call it sad, Dan,

28:49Mads Jensen:but it's just pure, cold-hearted, like commercial reality. I know, but we've got to be humans as well. You'll know who'd be sadder than you, Dan. Rishi Sunak, you know, who launched the AI Safety Summit in one of his biggest legacies as a prime minister was launching. You remember at Bletchley Park two years ago. I do. I remember it well. What happened to that thing? Does that even exist anymore? I think it moved to Paris. I have no idea, but that's...

29:12Lomax Ward:Nothing to see here. move along please no nothing to see here yeah yeah well so poor rishi is like you know having

29:19Mads Jensen:his legacy eaten away at again anyway so i'm not i'm not i'm just not surprised no on very quickly

29:27Lomax Ward:on your mads you raised the um anthropic accusing chinese labs of distillation so anthropic identified 24 000 fraudulent accounts and 16 million exchanges used by deep seek moonshot and minimax the chinese open source models to train their own models so there is industrial scale distillation going on which kind of makes me think well that's possibly how it's just gonna roll for the americans spend all this money the chinese say thank you very much they take all these amazing toolkits and build their own open source models from it so i do one quick question for you mads is at what point is that a national security issue i know trump's going to beijing soon isn't I mean, at what point is that a conversation piece between him and Xi Jinping, or will it just be left, do you think?

30:16Dan Bowyer:I think IP theft is always a national security issue. Yeah. I mean, stealing somebody's IP is no different from stealing goods, right? If you walk into another man's store and take his stuff and walk out, I mean, that's a huge issue. If you walked into another country and you start to take their assets and steal them and take them away, that's a huge, huge thing. And when it becomes big enough, yes, it becomes a national security issue. So, of course, it's an issue. The question is what you do about it. When you roll over, when Trump roll over just to try and do a deal? Well, it depends on what the deal is.

30:49Dan Bowyer:But I think with some of the stuff, some spillage is going to happen. The question is how do you create a system where there's enough mutual respect that you can keep it, you know, you can manage it. If it's all one way, then, yes, it's a huge issue. If you have massive global trade imbalances, but it's much easier to steal IP than to steal goods. So if the people that are making the goods, they get paid in hard cash, and the people that are making the IP have their IP stolen, that is going to create global destabilization. That's not going to work.

31:25Lomax Ward:I just wonder, as Anthropic on flavor of the month, positive flavor of the month in the White House, I wonder what kind of conversation they'll be having behind the scenes that will make any difference to Trump. And it looks like, I mean, correct me if I'm wrong, Jens, but it looks like the overall tariff rate with China has come down. So it feels like Trump is almost on a little bit of a back foot and trying to make nice and going and doing state visits.

31:52Dan Bowyer:Well, the problem was they figured out that China has got such a chokehold on the supply chain for U.S. defense. Yeah.

32:02Lomax Ward:So I don't know. Yeah, I guess it will just kind of maybe be ignored for now. But listen, on the more positive side, Anthropic Enterprise, same week as all of the challenges that we've seen. but um and so we enterprise they are now looking at integrations for salesforce slack intuit docusign legal zoom etc etc etc benioff and amaday appear together talking about claude for agent force 360 this all sounds like i mean if to me it feels very short-termist it feels like the question to the room is is anthropic now a partner rather than a destroyer i think it's entry point to kill it all off but i don't know mads have you got any any views on this one well there's always been

32:45Dan Bowyer:this discussion whenever you've got a new technology paradigm you've got the infrastructure players you've got the people that are creating the foundation are they going to be able to swallow everything and sometimes they do and oftentimes they they don't and companies have been trying to do this before ibm tried to build an application business but they found it was really hard microsoft has got a decent sized application business but still even when microsoft's grasp of the global tech ecosystem. They have not been able to become this application provider of choice for everybody. And I don't think Anthropic will either.

33:19Dan Bowyer:No single company will be able to provide all applications for all industries. So the question is, where's the margin going to end up? And there will be a huge battle. There will be lots of things to be reorganized. But I do think, fast forward five years, you will still see application companies. Might not be the ones we see today. Might be different companies. and there should be a lot of AI embedded.

33:41Lomax Ward:Or people build their own. I mean, I can't see this really happening, as you've mentioned before. Or there's just a system of record, build what you need on top.

33:50Dan Bowyer:Yeah, it turns out that building enterprise applications is a lot harder than a bit of Vibe coding, as we all know. We can build our own tools and we can do that in an afternoon and it's amazing. But then trying to roll that out to a company with tens of thousands of employees, with permissions, with data security, it's just a different kettle of fish.

34:09Lomax Ward:Yeah. Lennox, what else?

34:10Mads Jensen:Well, I just think stepping back, this is, I think, and maybe we're going to talk about this a little bit later in the pod, is like people were in such an era of change that people do not know which way things are going to go, which is right. And people are running around like headless chickens and you're seeing that in the market. And then in one week you have SaaSpocalypse, all SaaS companies are dead. then you have a bit of a love in between Anthropic and the SaaS companies. And everyone's like, actually, agents are going to unlock a tons more value for SaaS companies. And the SaaS companies that, okay, ignoring like very thin, like SaaS companies that don't have like great data, SaaS companies are going to be redeemed and saved by agents now.

34:59Mads Jensen:I think people just don't know where to go. People don't know which way it's going to go. So I think that's, to be honest, but it's what we talk about on this pod. It's what we talk about with founders every day. It's what we talk about with investors in our funds every day, right? People just don't know where it's going because it's a very exciting time to be alive. But that's true. We had SaaS apocalypse. Now we have SaaS companies redeemed and everyone has this bull case for how agents are going to absolutely improve and soup up existing SaaS products, right? It's crazy,

35:30Lomax Ward:literally crazy yeah it is it is kind of what it's funny watching the markets react to certain news or research pieces and we're going to talk about the people you basically got this book you

35:41Mads Jensen:basically got this bull this kind of thesis which is right so sas is a system of record ai is a system of action and this is this is great you know you need the systems of record and then you have the systems of action on top so um you know the winners are the sas companies that expose apis cleanly control the data layer, owner compliance relationship and price. Well, depends how you price, maybe even price on value rather than on seats. But isn't that, I don't know, this is the current thesis that people have. You know, we're now going from are SaaS companies cheap to are they expensive in the public markets?

36:17Mads Jensen:Because people just don't know. But it's exciting.

36:20Lomax Ward:We will watch it run from side to side. A little bit on AI margins and cost base, a few anecdotal aspects and then chime in if it rings any bells. So OpenAI gross margins fell from 40 % to 33 and 25, missing its forecast of 46%. Anthropic margins improved 40%, but still 10 % short of its own target. OpenAI inference costs quadrupled of their nearly a billion users, only 5 % pay and nearly half of all inference costs, which is nearly$4 billion, went on free users so i don't know if we want to dig into this at all but how do we see low max how is this going to play out how is this kind of margin piece and and the money actually going to play out do you

37:07Mads Jensen:think well i i think we're in the same i mean the same territory as to what we just talked about which is that people just don't know you know so open ai and anthropic as you said missed their own margin forecast last year you know open ai targeted 46 gross margin actually netted out 33 and Anthropic targeted 40 and ended up at 30. So, you know, both companies basically ended up at 30, 33 % gross margin last year. Burning cash hand over fist. I think this has been one of the questions of the last, you know, two, three years is where are they going to net out on gross margin? The answer is we don't really know.

37:46Mads Jensen:I mean, they've improved a lot. Anthropic's gross margin in 2024 was like minus 94%, right? So it's already improved a lot. And if it carries on on that straight line, then it's to the moon. You know, I think just in terms of the initial reaction to that news is first, you know, AI compute economics are tougher than we expected, right? Secondly, as you alluded to, free users are expensive. You know, that's roughly free users account for roughly half of open AI's inference. Thirdly, maybe they were probably a little bit ambitious, potentially in the context of their own fundraising. But one thing this will do, and this actually probably ties into what we talked about earlier in the discussion today around, well, people already know that the chips are going to, you know, the data center is going to get billed.

38:36Mads Jensen:The chips are going to get sold. They're looking at the revenue of like the AI companies. They're also going to be looking at the margins of the AI companies, right? And actually, this news probably pushes back the profitability of when do these companies actually get to break even, right? I don't know, Mads, I think Anthropic was at one point forecasting break even in 27. No, I think that's gone back to 28. OpenAI is more like 2030. But, you know, the longer those, so there's revenue numbers and revenue growth. And there's also kind of break even forecasting. and the moment that if they continue to soften, that is bad news for everything that's going on behind it.

39:18Mads Jensen:Well, maybe it doesn't matter for now.

39:21Lomax Ward:Maybe this is just kind of keep growing, keep pushing. Maybe it doesn't matter for now.

39:24Mads Jensen:You're right, Dan. You're not wrong, but at some point it's going to start to matter.

39:32Lomax Ward:Yeah, yeah. And also it's push and pull, isn't it? You've got NVIDIA at the top of the tree and then you've got the application layers underneath all of these incredible AI companies and all these foundational models. So that's got to, at some point, all connect the dots together. I don't know. Mads, any thoughts on this before we move on?

39:50Mads Jensen:You can't operate a giant tech company with a 30 % gross margin. What's that going to get valued at? What's the multiple on that? It doesn't work.

39:58Dan Bowyer:But as we know, next generation chips from NVIDIA will already have, per token price, that's a tenfold improvement over the current model. So we're expecting massive improvements there. I'd say to your point, I think you're right, Lomax, but everything depends on the growth rate, right? Last year, Anthropic grew nearly 10 times, 10x. Oh, for small base. Wow, it was a billion.

40:27Mads Jensen:Yeah, a lot. Sorry, I guess a small base in the context of big tech.

40:32Dan Bowyer:I know you have a lot of portfolio companies in your portfolio that grow from one to 10 in a year. But for us mere mortals, that seems like a reasonable growth rate.

40:44Lomax Ward:Well, thank you so much for the segue into chips, because I want to talk about European chips and I want to talk about European quantum. I guess the broad question I'm going to give you a bit of context is, can, should Europe double down on chips and quantum? There's been a lot of movement and a lot of activity in this space over the last year, quite rightly so. European quantum funding hit 1.5 billion in 25, up 170 % year on year. 11 billion in EU public funding committed over the next five years. Quantum Nation Fund 2, 220 million. I said this is a billion last year, but it's completely wrong.

41:18Lomax Ward:It's only 220 million, but it's double Fund 1. 55 North, 300 million quantum fund. There's a startup, Pascal and Paris, in talks to raise 200 million and a billion from a Nobel Prize winning physicist. There are two European hardware stories this week. We got one Dutch listing, Accelera, and there is the Finnish quantum computing company IQM that is going public via a SPAC. So it feels like we got some chips, we got some quantum. So this is all good news for Europe. Is quantum genuinely Europe's competitive advantage? can and should we do more in chips i kind of know the answer to this already but lomax can you kick

42:00Mads Jensen:this one off for us i'm going to cover quantum i know quantum obviously quantum computers have chips too but let's just leave chips for classical computing to one side for the moment which we can talk about a lot today and talk about quantum i do think we have a slight obsession around trying to find like europe's kind of the place where europe will be the best i think it's my obsession

42:22Lomax Ward:it is my every week i raise the same spectre than i was like can i can i lower the bar let's find a

42:28Mads Jensen:place where europe can be like not quite as shit right no i'm not having it and it might be quantum it might be quantum for a few reasons but don't forget i mean there's a lot there's lots of very interesting things going on in quantum in europe but there's also lots of things very interesting things going on in quantum in the us in china obviously and in australia of all places but But let's focus on, yeah, yes, indeed. Maybe let's focus on Europe for now and why, maybe let's start sort of to what the ingredients that Europe has, right? So Europe is absurdly strong at the hard science layer, right?

43:02Mads Jensen:You know, we have world-class universities and institutes, Max Planck, CNRS, Oxford, Delft, ETH, and Zurich, right? We have deep technical benches in trapped ions, neutral atoms, superconducting physics, cryogenics, quantum control theory. And this is why companies like IQM, which you mentioned out of Finland, Pascal out of France, Alice and Bob, also out of France, Universal Quantum in the UK, proud investor, even exist at the frontier, right? So that's number one. Like we have a really good hard science layer. The business world, look at ASML. ASML is like a unique company in the world. And actually going back to China and why China may be behind is because they can't get access to EUV, extreme ultraviolet photography, right?

43:46Mads Jensen:Which is what ASML do. So we know that's true. Number two, Europe is not, this is an interesting strategic difference to the US rather than like a kind of technical one is Europe is not betting on a single horse. In quantum computing, don't forget, we're still in something like the 70s when it comes to like classical computing. We are still a long way from, you know, having a kind of viable, useful quantum computer that can be used every day. And so we're still trying to figure out which architecture to use. Now, Europe actually has credible teams across the key and core architectures. Superconducting qubits, you have IQM, we just mentioned, neutral atoms, Pascal, error-corrected logical qubits, Alice and Bob, weirdly named French company, photonics, photonics, and actually trapped iron, universal quantum, again, very proud investor.

44:42So that's great.

44:43Mads Jensen:And actually the US is a little bit more all in on superconducting, right? And maybe they're right, which is maybe great, but maybe they're wrong. One other thing that I've noticed, the third thing is that Europe has done really well at actually using government money to fund quantum computers and become the first buyers of quantum computers, right? The Germans bought a quantum computer from the company I invested in Universal Quantum in the UK for 67 million euros, which is an amazing thing for a seed stage startup. And there's a bunch of other of those deals that have gone. If you look at IQM, which is this company now that's doing a SPAC, they've also got existing revenue from governments, right?

45:21Mads Jensen:So the European government, actually, something we criticize European governments for a bit here is they're not actually the first contractor of record for these startups. Well, first buyer enough is that they've been doing that, which is great. I mean, the Americans have been doing it too, but I think the Europeans have been doing it actually to quite a large extent. not the bare case, but the kind of slightly depressing thing as always comes back to the same thing. It's money. It's capital. You talked about$1.5 billion being invested in European quantum last year. Cyquantum, which is a US quantum company that originated in the UK, but whatever, raised a$1 billion round last year.

45:59Mads Jensen:There was something like$4.5 to$5 billion raised into quantum in the US last year. So$1.5 billion sounds like a lot, but it's not enough. if Europe wants to succeed. So I think to answer your question, I think Europe is really actually potentially almost best in class in quantum, has amazing teams, I think is at risk of losing some of them. Because if they start SPACing, they're going to SPAC in the US and they might start moving their headquarters over there. If people can pull their fingers out and get some more kind of 200, 300, 500 million, $1 billion rounds into these companies, then we might have a shot.

46:35Okay.

46:36Lomax Ward:I'll go with that. Yeah, I mean, IQM is listing on the NASDAQ, so we will lose it. But Mads, you're not convinced that this is the right frame and I shouldn't be so focused on what Europe can win in, right?

46:51Dan Bowyer:No, I mean, it seems to me that should Europe focus on one thing is a bit of a middle power question. And look, I mean, Europe is one of the three global economic blocks that have heavy weight. You know, we've got 16 trillion euros in pension capital. Today, we're exporting a lot of it to the US.

47:11Mads Jensen:But, you know, what if we actually redirected...

47:13Dan Bowyer:I thought it was double that, Mads. I thought it was in the 30s. I mean, it's possibly higher if you add in the UK, if you add in everybody else around it. It's a massive number. Yeah. And so people sometimes say, look, we can't be the best at every... Yeah. It's actual money. Yeah, UK has 3 trillion. Yeah, exactly. Exactly. Now... People say we can't compete at everything all at once, but people sometimes forget, 25 years ago, China was a developing country. And today they're competing across the board. I mean, 25 years, it's not nothing, but it's also not like 100 years. It's something you could decide to do and then go after it.

47:51Dan Bowyer:We start from a vastly stronger industrial base, ASML we've talked about. You've got Airbus, which is the global leader in aerospace. You've got all the quantum technology. the deficit seems to be a lack of ambition and incorrect capital allocation processes. So the plumbing of the capital markets rather than capability. And it seems to me that's entirely within our power to fix, right? You're sitting with all the Lego blocks here. We're just going to put together in a slightly different way. And I think if you take quantum, we have great teams, as I said. A hundred percent, a hundred percent.

48:24Dan Bowyer:So that's my take on it. I do think we've tried with the CHIPS Act to drive more semiconductor manufacturing into Europe, and it's failed because we're partnering with Intel for the sophisticated chips, and Intel is in trouble, and we partner with ASML for the less sophisticated chips, but of course they're the only ones who can really make the sophisticated ones. And it seems to me that maybe we just need to be a little bit more, should we say, aggressive? Yeah, you could say that in our negotiation. Now, ASML is a fantastic company. Maurice Chang is one of the entrepreneurs and founders of the last 100 years.

49:05Dan Bowyer:He's incredible. Love the business. TSMC relies on ASML equipment to do the things they do. Why is it too much to ask for that they put some of their sophisticated plants in Europe? It seems to me to be an entirely fair quid pro quo. We know that from a supply chain perspective, everything that's in Taiwan is at risk. It is our entire economy is dependent on something about this.

49:31Mads Jensen:Samsung, you build your fab here. TSMC, you build your fab here.

49:35Dan Bowyer:Right? That's happening. And I just think we should do the same thing.

49:39Mads Jensen:And we shouldn't be shy about it. Did you see, and this is not a party political broadcast, but you know these guys... I wonder where you're going to go now. You know these dudes reform. Reform was saying this week that they want to basically co-opt the roughly 400 billion pounds in local government pension schemes and they're now kind of managing a disparate way, pull them together into one big pot and basically force them to invest into PK businesses. And nationalized pensions.

50:05Lomax Ward:Is that part of their manifesto?

50:07Mads Jensen:Well, it's not nationalized pensions. It's what we've been encouraging Rachel Reeves to do and what she's been trying to encourage the pension trustees to do, which is stop investing in US bonds or maybe US equities possibly, but actually start investing in local private equity or venture capital, clearly not too much of a pension into that, but there should be a lot more than we have now.

50:29Dan Bowyer:How does it make you feel on a personal level when the party you're most closely aligned with on a capital markets perspective is reform?

50:37Lomax Ward:I've got to say, I read the UKIP policy when UKIP first started and obviously everyone's going rage on the racists. I read their manifesto cover to cover and a lot of it was extremely sensible and thoughtful. It wasn't the ragey, what I was expecting to be like, throw out all the brown people oh my god we're gonna do this and you know the daily mail is gonna be our mouthpiece it was it was pretty sensible smart when you've basically got especially

51:05Mads Jensen:when you've got an incumbent government that are basically launching business prevention policies left right and center you know and actually the bar is quite low to create sensible coherent you know investment and business supportive policies no i'm sure the conservatives will

51:22Lomax Ward:have a go but actually but they screwed the last 15 years before labor though i mean it was all right up until 2016 wasn't it uh yeah maybe i i it just looks it all we get the we get the governments that we deserve and until they're a smart business people in there who understand how business works i know rishi was possibly our great last hope but unless you have smart business people in office who understand these things i don't know i don't know how you do it i mean you can't have people who've never had a job in their life running a department which is making decisions based on how business be careful what you wish for you get you get you get you know

52:02Mads Jensen:you get trump who wants to do a deal you know a deal a minute would it be would it be alan sugar

52:08Lomax Ward:who do we have as pm i don't even know i mean matt clifford of course yeah yeah the last thing This will possibly be a very quick section because I know that it is neither of you are happy place. But I want to just quickly talk about the Citrini, the research piece that was done by James Van Geelen and Al-Apshar, I think it was. And it was roughly titled The Consequences of Abundant Intelligence. Now, it is a research piece. It is just effectively a thought experiment posing the statement, what if the AI bulls are right about AI capability? And what if that very success triggers a macro crisis?

52:52Lomax Ward:And then they go into the mechanism is effectively intelligence displacement spiral. So what they're saying is AI improves, white collar layoffs, spending collapses, consumer economy deteriorates. But AI CapEx continues because it's now the OPEX substitution, not discretionary spend. So we start losing our brains to AI and the systemic cycle completes. Listen, I know that a lot of this do me stuff is very easy to talk about. it's probably not going to happen in this way, but you've both read the report. Mads, and I know you hate this kind of stuff, but is there anything that you would pull out on either side of the case that you think is interesting?

53:37Lomax Ward:And I fully understand if it's a very, very short segment.

53:40Dan Bowyer:No, I mean, let's take it at face value. So the thesis is exactly as you're saying, the doomsday scenario. AI is going to displace all the white collar workers and therefore they don't have any money anymore because they've all been fired and therefore they can't buy anything and demand collapses and we're all going to hell in a handbasket. And isn't it just terrible? Now, the steel man of this is that because AI automates cognition itself, so the process of thinking, there's no more work for us to do. AI will do everything. So there's nowhere to redeploy the workers. Even historically, it's always been one kind of job disappears, new business types show up and people go there instead.

54:20Dan Bowyer:But AI is just going to do everything. For me, it just does not seem credible. There is so much work to do. And if you think about two vectors to this, one is the capability and then the rate of deployment. Now, the world is not going to change in three or four years. We're not going to go from everybody's employed to nobody's employed. That's just institutions are far slower at adopting and changing. Another thing to think about is we have never been able to predict the new types of jobs before. Social media manager was gibberish in 1995. And so have all jobs been for every single technological revolution that's gone before us.

55:03Dan Bowyer:So we've always found new things to do. In so many areas, humans want things from other humans. The relationship is the product. You go to see a doctor, you want somebody to hold your hand and tell you everything's going to be all right, not just talk to a robot. We want teachers who can help be role models for our children. We want the human contact. And so I think there'll be plenty of things for us to do. But there's no doubt, and this is possibly where the article has a point, is that there will be disruption. And the question is how we deal with that disruption in the intervening period.

55:41Lomax Ward:I think it's worth a read. I always liked these thought experiments just for its very own sake. I don't believe in the doomerism. I don't believe in the accelerationist viewpoint either. I just love the thought experiments and people have taken time. What was the other report? Was it the 2027 report? I forgot the name of it. Where they've set out a thesis over a period of time. And obviously this Cetrini research where they basically posited out two years in advance and then done a press release as if it is 2028. I think they're just really, really interesting thought experiments to get my noodle going.

56:19Dan Bowyer:So just linking in something else, because Jamie Dimon, he's obviously an incredible business leader, and he's sort of been out talking about what's happening right now with Caspocalypse and kind of what's happening in markets. We have seen that Blue Owl, they took a big hit. It's a private credit company that went public having lent a lot of money to SaaS companies kind of against recurring revenue that now looks like it might not be as recurring forever in perpetuity as people thought. And so what Jamie Dimon came out to say was, look, there's always a big surprise that triggers the credit crisis.

57:00Dan Bowyer:We had it with subprime in 2008 and people didn't think that subprime, it came out of nowhere. That's why the big short is such a great read. The whole thing was just it took everybody by surprise. And Diamond's point is maybe software is the surprise that's going to create a credit collapse, right? Or at least a run on things. It doesn't mean that, you know, clearly, you know, after 2008, you fast forward a couple of years. I mean, the economy is sort of back to New Calipari. It doesn't mean everybody loses their job. But it could be the thing that triggers in the short to medium term some reallocation.

57:35Lomax Ward:Yeah, it won't be much, will it? And I think I'm right in saying that every crisis we've seen has been credit driven. So maybe there's maybe it'll be software this time. Maybe it'll be another credit crisis. We'll see. I've got one deal of the week and it's a super short one. I just want to wave to wave who are raising 1.2 billion at 8.6 billion. They've raised 2.5 billion to date. Now, I think the only investor in this round that I've not seen before is Nissan. It's their first ever automotive industry investment in Waverton. He's great. I love it. Wave, well done. You're cracking it. Go, go, go.

58:16Lomax Ward:Anything else on your mind, Mads, before we clock off?

58:20Dan Bowyer:What's happening in the week ahead? We've got a UK spring statement coming up soon. we've got possibly a deep seek model coming out that's been rumored here for i thought that was this week this is deep seek four isn't it i thought that was that was happening i don't think we've seen it yet and so there's and there's the open ai 100 billion dollar round of course is that is that going to close here by the end of the month and then of course as we spoke about earlier on the hexeth ammo day showdown uh is anthropic in a cave or not so yeah there's a few things to look

58:54Lomax Ward:out for i think i think i think it's a big fat yes but we'll see we'll see how much uh dario sticks his heels in but i imagine it'll be yes please thank you so much sir we'd like your money but we'll see my dear men lovely always a pleasure never a chore catch you next week see you in the

59:09Dan Bowyer:next one