E691 | This Week in European Tech with Dan, Mads and Lomax

27:56Biziou, exactly. Which I think was the name of like some, a Microsoft product back in the day. It is a Microsoft product, yeah. That I think has been mothballed. So anyway. Watch for the legal proceedings, then we'll see. So this is, they've said, everyone, you've got to be using this by 2027. This was developed by DINUM, the Direction Interministerielle du Numerique, which is the French government. Very, very posh. Loving the accent, darling. Digital affairs. The reason, the breaking point, and this is something we've discussed a little bit before here around sovereignty of tech platforms. But I think the breaking point here was a recent incident involving the International Criminal Court, you know, not Donald Trump's favorite organization.

28:37when the US issued sanctions against one or two of the judges, I think, and Microsoft reportedly cut off their email access. So this was the, I guess, the straw that broke the camel's back. I mean, look, I don't, you know, I don't even know why we're taking this sort of, do we really think that a French like state developed software product is going to displace widely what the best of the US, the American like global tech can produce? I don't think so. So don't forget, this is not the first time in the context of France they tried to do this with cloud. And they had two competing cloud providers in the late noughties, 2009, 2010, which both kind of petered out.

29:17This is happening. It's not just France. In Schleswig-Holstein, in Germany, they migrated 40 ,000 state workers from Microsoft to open source alternative for their inboxes. And also the Austrian army have done this as well. so i think this is a sovereignty push by the you know governments in europe i think it's going to be a lot harder for them to do this by mandating and creating their own products i just you know unfortunately you know as a free market libertarian or certainly if you believe in the ability of the free market to create the best products for the best users etc then this is not the way to go one counter to that loan that is the technology i mean this isn't the noughties This is technology is so functionally smart, capable, ubiquitous.

30:05There is a, it's a bit of a step change. AI obviously layered on top of that, but it's a bit of a step change technically. No. So I think it makes it easier now, if you are this denim or whatever it is in France to actually create, you know, a new zoom. It makes it a lot easier now than it was 10 years ago. Right. So that is for sure. but again like i tend to be more of the the in the camp of you know you let the you know you let the market create the best products if you want to create national champions there are other ways to do that which is to create great environments for entrepreneurs who are you know able to raise money you're not weighed down by a bunch of red tape can raise growth capital can ideally win contracts with you, the French state, as an early user.

30:55So I don't know, that's the kind of model that I prefer, rather than being like, we're going to fork this piece of code internally in the French government, we're going to kind of put some, you know, our own system, our own skin on it. And this will be what everyone's got to use. I just don't think that's the way to go. I 100 % agree. We have incredible founders in Europe, and we got it like, what, why did we need to, you know, why do we have If we needed to chase some of them away, right? We found our Revolut, we chased off to the Gulf. You've got even things like Telegram. Telegram could be an alternative to WhatsApp.

31:34There's so much we could do. When it comes to technology and governments, this is actually, it feels like a missed opportunity. Linux was invented in Europe. There's no good reason why European governments don't use it more. That feels like lazy procurement. It's an incredible software today. You could use it to run operating systems. You could use it, you know, open source alternatives to the Microsoft suite. So I think there's good things in the market you could do today. And then you need to support founders instead of sort of trying to do, you know, concrete, kind of heavy-handed government built things.

32:06I think Saul Klein said this in an FT article when asked about this. It feels like we're fighting yesterday's battles. Oh, always. You know what I mean? Always. One thing it's important to remember is that the European Union exports 500 billion euros of stuff to the US every year. And obviously, we'd like to keep selling those things. That's quite important for our economy. So, yeah, it feels like we need to tread carefully and not too heavy handed. I was chatting to Gemini about this and it came up with the it feels like a kind of digital version of the Maginot line. You know, as in like, you know, it looks great in theory.

32:47It looks amazing, but, you know, it ends up being totally useless when it really needed. So I saw Pedro Sanchez getting serious about social media and misinformation, disinformation, hate speech. Well, he's now been called Dirty Sanchez by Elon Musk. Elon, Elon Wade straight in, obviously. I don't know where he finds the time to have all these have all these have all these pissing matches but I really enjoyed watching a world leader I got that same warmth when I was watching Mark Carney as when I was watching Pedro Sanchez basically saying that he's going to hold CEOs accountable and you can't hide behind these tech platforms anymore and he's making hate speech and other forms of mis and disinformation online a criminal offense now it does then and obviously Also, he's removing the ability for under-16s to use social media, under stuff that we've spoken about before.

33:43But it does then raise the challenge and the question of how are you going to prosecute? Where are the free speech lines? And how is the state going to control? And that's really, really simple. It's a very interesting one. I would say an observation is that, and actually it's quite a smart one by the European leaders, is they turn this into a public health issue rather than a free speech issue. So before it was like, we want to control what's being said on the platforms, which there's always going to be debate around that. And the Americans say, well, you're expressing free speech. And then the Europeans say, no, this is hate speech.

34:17But actually now by saying, look, this is actually, I think Macron said, this is damaging the brains of our youth. Right. Literally came out and said that in the context of social media and children. It's like, well, actually, this is this is being treated as a public health issue, takes it out of the ambit of that debate on free speech. and I think it's a very smart thing for them to do. And this is one area where I've said before that I do agree with them. Even in principle, I know it's going to be hard to implement these things. But everyone in Europe now is moving to this. It's a slippery slope to censorship?

34:51No. I'm just thinking back to COVID, how so many things couldn't be said and how much censorship there was at the time, under the guise that this is for the purpose of public health. Yeah. No, I guess I'm an optimist and I always see the good in people and I don't think that's right. But you're right. I suppose that is a risk, yeah. But I think when you're trying to, I mean, maybe we're kind of conflating two separate issues here, which is the general kind of free speech debate and in the context of under-16s, like preventing them from accessing social media, right? I don't know. I'm actually, I don't think that's a super-estope.

35:31Yeah, no, that's fair. That's fair. I think we were sort of talking about, you know, what are you allowed to say online and what should you be allowed to say? I think it's different. I mean, we also don't, we also, there are also other things we don't let children do, right? Should Donald Trump be able to call, you know, Pedro Sanchez, Dirty Sanchez? Of course he should, you know, if he wants, you know? But that's not really hate speech, but there are. Well, true, but it's offensive. It's offensive to some people, but yes. Yeah, but I mean, offense is a choice, isn't it? I mean, now you're getting into the debate, but yeah.

36:01Yeah, maybe we can't, but we'll see who they prosecute and for what. So we will soon hopefully see some of the canaries as first in the coal mine as far as legal recourse goes. So I guess we'll see. We'll see what he does. Well, and also don't say, you know, the French authorities have been raiding the offices of X this week. That's coming. There's now processes being kicked off in the UK. So the regulators are now really going against X in the context of this, you know, pornography kind of line of thread. Well, I just wonder how much they're trying to use that as a wedge. And this is what worries me is that this becomes about much more because of course, you know, X have been doing things they should not do that are inexcusable.

36:49But there are also a lot of debates It's about what should you be allowed to say. And I do feel when it comes to freedom of speech, I think that's a very important principle. But what would you do if you were the UK government, Mads? What would you do about Grok? Unpleasant image creation and unpleasant messages disseminated via Grok. What would you do? Can we not find them? I guess that's where it will end. But you need the evidence, right? So, listen, who knows what's going on as part of a criminal project? Is that all it's about? Is it just about gathering evidence for that fine? I mean, they're now merging with SpaceX.

37:25I think you can fine up to 10 % of global turnover. I mean, it's going to be a pretty meaningful number. And then watch it stuck in the courts for the next 10 years. Is everything just getting bogged down until people get bored or run out of money? Well, a lot of the fines that were made, I don't know if you remember, there were kind of all the issues with companies escaping to Ireland and not paying tax there. and the money eventually was clawed back and it was tens of billions. Was it? I always wonder what happens at the end of these fines. It always looks like it just becomes noise and then nothing.

37:57I always wonder where they settled. Listen, I think, you know, Musk, he is exhibit A in the discussion around the corruption of power because, you know, power corrupts and absolute power corrupts absolutely. And it's not so much about greed. It's more about unaccountability that corrodes clarity and claritive thinking. I think in many ways, I think he is still mission driven. I think a lot of the things he does things because he believed them as opposed to caring about whether he has another billion in his bank account. I think he stopped caring about that a long time ago. So I think the world needs mission driven entrepreneurs, but they also need people to be accountable to the law.

38:34And the question for me is, can Europe become a place that welcome the best entrepreneurs and hold them accountable? I think all this talk about criminal liability and putting people in jail and all these raids, I think it's a bit of a slippery slope. I think it's dangerous. Maybe it's the right thing to do. I don't know. But I just, I think we should be really, really sensitive to it. Kind of the heavy handed, heavy handedness of politicians jailing people because they don't like them or because they don't like what they say or because they feel they're a threat to power. I think that's very, very tricky.

39:04But fines can't be the way when you've got more money than God. I mean, almost to the point of it being a rounding error to them why would he care well i mean look if the number is big enough eventually people will care i think i think you have to hit them where hit them where the ego lives you have to you have to kind of you have to they have to they can't they can't operate in this territory there's some there's some form or version of control i can't see a fine being the answer but we'll see in fact thank you so much for the beautiful segue into muskonomics because this has to be my favorite topic this week.

39:38I've been looking at data centers in space and trying to understand more of this$1.5 trillion combined merge. And I can't make any of it make sense. The physics, the math, the payload logistics, the science in space. I cannot see how any of this story makes any kind of sense so mads set this one up because i really want to dig more into this but i mean i think i think you're you're exactly right dan i mean it's exactly tied to what we just spoke about the success breeds hubris he's one of the most impressive entrepreneurs of our generation and you know i need a hubris hubris proceeds nemesis no well there you go right icarus flying too close to the sun which which might literally be what's what's happening here But I think his track record in that sense is not in question.

40:29He's done incredible things. But it's just that that success has led to a hubris that's eroding clarity. And it seems clear that XAI just can't catch up from a commercial perspective with the anthropics and open AIs of this world. They're trying hard, though, Mads, aren't they? They're spending a billion a month to try. They're spending the money, but they don't have the revenue. And so they've got to find it somewhere. And so he is merging one company controls into another company controls, setting his own price. You know, no board, no checks. And it just it's wrong in so many ways. And I think it's a shame.

41:08It's a tough gig. No, it's being wrapped up in this kind of broader Muskian narrative of, you know, data centers in space. like all of this is interconnected you know like space ai compute data centers we're going to run i can buy the full stack when i first looked at it i thought no i can absolutely buy this kind of full stack method this thinking around that that makes complete sense but then when you look at the detail it makes no sense the irony here also is that don't forget about a year ago that there was there was basically a plan to demerge Starlink out of SpaceX to actually, you know, unbundle these organizations, right?

41:50Because don't, you know, remember that Starlink is the kind of cash cow within, you know, the launch business at SpaceX is a really a very valuable business because it does, you know, 70, 80 % of launches globally, but it, you know, it doesn't make such money. Whereas the, you know, Starlink is printing money. Let's remember where some of this has come from. So, you know, you go back a couple of years, there were huge discussions about alignment and AI is one thing, but having AI that's politically correct is another thing, right? Having Google's Gemini show that, I don't know, George Washington was black and, you know, Nazis were Native American women.

42:27It was just diversity craziness. And I think Musk having some more of a free speech bent and sort of say, listen, we've got these AI models that we have all this stuff that's built in and they're not truth-seeking. They are something else-seeking that certainly is anything but the truth. So I think the mission to try and create an AI that is maximally truth-seeking, I think is good. But I think the problem is he's gambled on trying to make it into a commercial endeavor. He's managed to raise some money. He's brought investors along with him, but he has failed to scale it because he's not focused enough on commercialization.

43:00And therefore he's doing the easy thing, which is actually taking cash from another company that he controls but doesn't hold he own. And from a corporate governance perspective, it's wrong in so many ways. Because, yeah, because basically SpaceX is making, SpaceX is finally profitable, right? It's doing like one to two billion, I think, a year, right, in profit. Obviously, XAI is burning, what, a billion a month. So it'll cover only a small bit of that, but it'll cover a bit of it. The revenue on XAI is negligible, right? It's 400 million or something. But SpaceX is one of the hottest, one of the hottest private companies on the planet.

43:34Everybody wants to buy secondaries. It looks like they're going to go public. They're going to be able to tap public markets. And of course, that's the thinking. We have a limitless ATM here. It's very interesting also that I love the data center and space narrative because, you know, the guy, the guy, Philip Johnston, who's the founder of StarCloud, which is this YC startup that has been that did its first test a couple of months ago. that's basically a seed stage company a year ago that has been a big exponent of daisy centers in space has like he's english by the way which i think is really cool even though the company's in redmond in the us but it's actually a british founder which is you know good to see but actually he's been amazing if you follow him on on socials at like beating the drum on this narrative that only really started to pick up in like q3 last year and then musk has kind of been you know beating that drum since then and now it's kind of convenient to tie the ai story into the whole space story now how does he talk about latency and payload delivery and cooling and just some of the basic physics about running data centers in space and you can not you speak you listen to elon he'll say oh you know it's okay you know the the chips they'll they'll they do bit flips there there is error correction where everything so far has been on the older chipsets where they're a little bit more resilient and you look at you look at 50 of anything that goes up into space fails and you go it's okay it only needs the last three to five years and you can recycle it but that's fine if you're looking to push a million units up into space that's like 10 launches a day and that's ignoring the fact that just the solar panels required just just the energy required look at the iss it's got like football fields size solar arrays to collect the amount of energy that's required it's like none of this makes sense from the cooling from the energy from the payload yeah but i'm going to go against you i'm going to go against you on this one or i don't know where you were going but say look if anyone is going to figure this out sorry i was just worried if anyone's going to figure that out it's musk no but it goes against and also he but no no no no If anyone's going to figure out, it'll be Musk.

45:52This is going to be profitable within three years. He's saying. Come on. Yeah, yeah, of course, of course, of course. But, you know, he does that the whole time. But, you know, Musk, like, he now has the full stack. He has, you know, because if you know, like, you know, this guy, Philip at StarCloud or any of the space companies I've invested in, right? It's very hard for us. You know, when you're building a space company, a component company or something that needs to prove, like, something that it can do something in space, you have to like book your flights, you know, eight to nine months in advance, right?

46:22And like you can do that one test flight, maybe two a year, depends on how much money you have, right? So for startups doing new things, they can maybe put something in space once a year, show that if it doesn't work, then you've lost 18 months of like runway in a startup, basically with a failed experiment, right? With Musk, he could be launching the whole time, running experiments at mass scale with, as Mads pointed out, the huge capital weight of the public markets behind it. I honestly think you'll figure this out. So that's the bear case for it, for sure. I think we are a decade away, even at Musk.

47:03Sorry, did I say bear case? I meant bull case. No, I heard you. I heard bull, don't worry. but I just can't see how any of those points can be easily. No, they're all massive points. It's decades of R &D to get this kid to work in space. Even at his rate of knots, there is no way that you can counter all of the failure, all of the weak points that are going to make data centers in space viable. It just feels like a PR story on steroids that makes no sense to me. But we'll see. Never bet against the man, right? That's what you're saying, Lomax. Never bet against the man. Well, I'm also just like, there's a need, there's a massive, like, if you could sort out the problems, like the cooling, the latency, et cetera.

47:48Latency, cooling, payload, energy. I'm not a space engineer. What the hell do I know? But it's, it just, even a base looks very, very tricky. Yeah, but imagine if actually you sorted it out. Could you imagine, like, you know, just look at the Europeans trying to build data centers now next to school playgrounds. There's no space. And then people won't let... It takes years for these things to get from you. Weak argument. Weak argument. I don't... I mean, especially... I mean, look at what Alphabet's doing. But Dan, if you follow the money, it makes sense, doesn't it? Go on. So, Musk wants to build XAI.

48:32He wants to have the world's best AI. He has, like everybody else, seen that AI is the most important technology of our lifetime. The only way for him to keep funding it is to find cash somewhere. And Tesla is struggling, right? They need to have their own things to sort out. X is the best way to get that limous ATM he needs. And so what argument does he need to create a reasonable business case that you can merge two companies without too many shareholder lawsuits? Well, there's got to be an argument. That's data centers in space. That's it. The data centers in space don't have to be real in the next three years.

49:10All he needs is the argument that they could be. Yeah. And then we're back in bubble territory. Oh, 100%. But don't forget what he did with Tesla. It was also a question of, well, just have faith, have faith, have faith. Right? Like with SpaceX. Have faith, have faith, have faith. But he delivered cars. SpaceX delivered payloads to the ISS rescuing astronauts from X and Y. So there was a proportion that worked. My thesis here is that data centers in space cannot work at core. Great PR, raise the money, sure, go public, I get it. Everyone's going to buy into Muskonomics. But there's not even a core version of reality.

49:49There's not even a car on the road to then stretch the truths around FSD. you haven't got details details details you're in you're in the wrong industry man you know going and listen i know i've got to think bigger and not be such a cynical old fart but this is just this is just one step on the crazy train too far i was having a chat with one of i probably can't disclose too much but a startup that we back that's working on propulsion and space the stuff that we talked about yesterday would blow your mind about what is possible and what is going to happen in the new space economy right and it's stuff that i came away from the call like bamboozled but also just like yeah actually this could happen and i can see why and you put enough smart people on the case and you put enough capital behind it it will happen so good let's put some money on the table what do you reckon he says three years his exact quote was no i think i don't 30 months to profitability from data centers in space.

50:55I'll take that bet every day of the week that it is not going to happen. Anyone up for that? I'd give him maybe five or six years. But, you know, it's still going to happen, Dan. All right. All right. We will see. Did you see the news this week that the Russians have been patrolling around space with two nefarious satellites? Tampering. Yeah. And it seems to be quite easy to take satellites out. And maybe that's part of the reason why he wants so many up there. Maybe it just makes it a little bit of a numbers game. But it does feel like it's relatively easy with lasers to take satellites out. So I don't know.

51:34Who knows? We are going to see space lasers after all. Reagan was right, baby. Yeah, they've been lurking. They've been lurking. Actually, one of them, I think, has just fragmented and blown up. So that's a good sign. That's one doubt. that's that's part of the course for the russian stuff last up but not but not least but a little bit of a funny not funny was the anthropic super bowl ads did you see them i thought they were kind of cool so they're basically poking fun at chat gpt for ads going back to our top of the show points and then sam a coming on x and going great funny funny but then having a bit of a dig and getting quite acidic and it's quite unusual i thought to see the ceo of a company actually front it so anthropic going i can't believe you're putting ads in ai we are never going to put ads in ai which is also an interesting stick in the sand for anthropic to take but um other gloves off is it why why are they going for each other why bother doing this anyone got any thoughts on this i think the uh the team over at anthropic thought it was too good an opportunity to pass up on they obviously have a spring in their step 2025 was a formidable year for them let's not forget They started last year at what seems like around a billion in annual recurring revenue and ended close to 10.

52:51She's never seen anything like that before. You come out of the year like that, you feel you can walk on water. And then for sure, somebody has said exactly, Dan, as you've done many times, well, aren't users just going to leave when the ads are switched on? And we know OpenAI, they've been playing around with the idea of putting ads on for a long time. They hire people out of Google's ad team. are they coming are they not and they told us recently they're coming the ads are coming and we know they have to because they gotta monetize all those users that aren't paying today and i think they just felt it was too good an opportunity to pass up on yeah and i guess actually thinking about it and listening to you mad it kind of makes sense because what they're basically saying to the market is chat gpt took the brand everybody when they talk about llms or ai immediately think about chat gpt because they were first out of the gate so they're now trying to and obviously anthropic then focused on the engineering and the coding and more of the enterprise and the back office.

53:46And now they're trying to go, look, we are back in the race. We are in the game. Here we are. It could backfire, but maybe it's an interesting brand play. Maybe just on that basis. There's no bad PR, right? As you know, I'm a huge Anthropic fanboy, but I actually, I didn't find them that funny. I thought they were a bit... They're based. Yeah, yeah, yeah, just a bit. Yeah, exactly. I thought they were hilarious. I thought that the dating one, when he's talking about advice or speaking to his mom, and then, you know, come and meet golden oldies and get in a date. I thought that was like a little bit spicy and a bit funny.

54:22I thought that was quite funny. Hey, listen, the difference between you and me. If I was the CEO, I wouldn't have greenlit it, but maybe they'll get away with it. Yeah. Anything else before we close off? Yeah, I think we should just talk about this critical minerals thing, no? Do it. So we had a reality check this week released by the European Court of Auditors. Must be a fun Christmas party there. But a reality check on the EU's Critical Raw Materials Act, which was passed in 2024, which set a goal that no more than 65 % of any strategic mineral should come from a single country by 2030. And the new report basically says the EU is set to fail that target across the board.

55:06Surprise! You know, we currently get, EU currently gets 97 % of its magnesium from China and China still controls 70 to 74 % of the six rare earths needed for permanent magnets. So, you know, way off. And then there's this amazing deposit found in Kiruna in Sweden, a rare earth find, but the report notes it takes 10 to 20 years to actually start mining in Europe due to red tape. But this is the problem, is it, Lamex? It's just an inertia problem. Rare earths aren't actually that rare from what I've read and from what I can gather. It's just the mining and the processing, which we've outsourced our dirty processing to China because they don't mind polluting their rivers and lakes.

55:52But we need to A, find it and then mine and process. But it's just that period of inertia. We don't have the infrastructure to do it. That seems to be the problem. Well, we don't have the infrastructure to process it at the moment, but all of this stuff could be built, but to permit it and find a place to do it is very, very hard. And that's 15 years, right? I mean, whichever way we cut it, that's decades away. It's a long way away. And actually, interestingly, just on this, on Rare Earths, is there was Rubio and Vance hosted a Rare Earth Summit in Washington, D.C. this week. Yeah. Which is basically the theme of the summit was, how do we, the rest of the world, apart from China, like wean ourselves off China, right?

56:31So it was like putting some investment into certain African countries. And the Europeans were really trying to cozy up to the Americans here to try and come up with a kind of pact and agreement. It's funny, all these things tie in together. We talk about the French coming up with their own sovereign tech, MADS, the 500 billion of goods that we send over there. All of this is interlinked. And there's only so much we can rattle the saber and be like, oh, you can't have access to our market for this. or we're going to shut down and use open source for our emails. Globalization works, right? Yeah, the under-16s can't use your social media platforms to look at your adverts, that kind of thing.

57:10But it is all intertwined. And then we need to go to them to be like, what are we going to do about rare arts? The great uncoupling can't uncouple. And there will be another shuffle for all these minerals and components that we need to build these bloody chips and this infrastructure and data centers and energy and sovereign control because we all need the the rare basic components to do all this stuff oh my goodness that's it i'm moving to mars are you know max you've got to do your deal of the week because this is all your baby drumroll finally ladies and gentlemen i am now talking my own book and this is my deal so i met a young portuguese founder in his 20s called pedro in early 2024, he was working on a problem that I knew about, which is clinical trials are incredibly expensive, right?

58:02So the cost of drugs, the cost to get a drug to market is anything from two to three billion now. And a large part of that component is because clinical trials are getting more expensive and more difficult to run. So he is basically taking LLMs to try and solve that problem and make clinical trial planning much more streamlined, quicker, and save a bunch of costs from avoiding protocol amendments during the trial, right? And he's had unbelievable traction. So we committed early in 24, seed round was done, you know, mid late 24, big rollout last year, commercially sales grew very fast and raised the second biggest series A in Iberian history, 52 and a half million dollar series A, which is a pretty big series A for Europe.

58:46I think it's top, top five by quantum for last year in the european series days so that company is on a tear and well done pedro and he's doing something very important which is cool and and i mean what a year and a half in so nice yeah yeah no i mean this is actually this yeah it's you know it yeah exactly it's very exciting well done pedro well done lomax i wanted to just raise five european unicorns so aikido cast the eye harmerton sapiens and pre-plug which is ukrainian edtech so i just wanted to highlight in january alone we saw five european unicorns and not not deal of the week but i just wanted to give them a shout out i think no that's good and aikido aikido is a hell of a company because aikido is basically you know they've gone from standing start to series b in a couple of years just rolling out in the us now i think it's just it's just important to uh important to highlight that.

59:41Mads, anything on your mind before we close off? Yeah, we often talk about how there is too little capital at the growth stages in Europe to fund some of these big, you know, successes when they scale up. And so great news that there's a new one billion euro fund that's launched here in Europe. It's not Karamba, it's Kambara. And so the idea is to deploy kind of checks exactly at the size you're talking about, Lomax, kind of the 30, 40, 50 million euros into great companies. It's a team coming out of firms like Mundi, Atomico, and Promus, and other kind of European institutions that have come together to create this new business.

1:00:21And they have announced they've already raised€750 million, so congratulations to them. Aye, caramba, not caramba. Well done to you. Boys and girls, a pleasure in a gift. We'll catch you next week. See you in the next one. Bye.

1:00:44Upside!