E711 | This Week in European Tech with Dan Bowyer, Mads Jensen & Lomax Ward

0:00Lomax Ward:Now, starting at the top, gents, the oil shock. We have to talk about this. So Trump, I think, has just kicked off a global recession. I just think we don't know it yet. Iran blocking the Straits of Hormuz for the past two weeks is a massive problem. Oil and gas coming through the Straits, as we know and have spoken about before, roughly one fifth of the world's supply comes through that strait. For context, I was looking at the numbers. That's about 20 million barrels of crude per day and about 10 billion cubic feet of LNG. And that really puts into context the 400 million barrels that was released by the IEA member countries, which is effectively around a 20 day buffer.

0:38Lomax Ward:So I think it's just not really going to matter all that much. The US we've seen has now reduced sanctions on Russia. So obviously they're going to buy some Russian oil. My take is the damage has already been done. Inflation up, interest rates steady or up, not good news for us. And I think we are in global recession territory. The other thing before we dig in, I wanted to just share is I was looking at the new leader of Iran, who's an incredibly strong military guy, unlike his dad, obviously devout and driven. He is not backing down after seeing his father killed, his wife, his sister, brother-in-law and their daughter.

1:12Lomax Ward:And by his own words, there will be revenge. This is not going to look good. This is bad news. I mean, he's extremely wealthy, powerful, military guy. He's got a personal fortune of around three billion. He's got about a hundred billion set aside outside of national wealth and control of parliament. And he knows how to use it.

1:40Lomax Ward:so mads a lot of word salad there pair it back to europe pair it back to us startups investing and all that good stuff yeah a bit of a pickle we're in i mean we were talking about the anniversary of the you know russians russia's invasion of ukraine a few weeks ago and you know as we'll recall at the time that meant that you know we could no longer rely on russian gas. And so we turned elsewhere to buy that core energy supply for the continent. And some of it we could get from the US, but a huge portion we got from Qatar that's no longer available. And this means massive impact to the European economy.

2:19We've seen the gas storage is 40 % down from two years ago. And we've seen the price of gas already doubling in sort of the local wholesale market. And that means that the diversification strategy we put in place after 22 is now itself broken. We've also seen that we have kind of economically, we have kind of multiple factors that are hitting our economy. On one hand, we are ramping up spending on defense. We're expanding the deficits because of that. And so that in itself is inflationary. And then you have this inflation that will hit from energy markets if this doesn't get resolved quickly. So that's a bad cocktail.

2:56could lead to higher interest rates. When they're higher interest rates, people are potentially less likely to allocate to alternatives. That means less money for venture capital, means less money for the startup ecosystem, means less growth. So it's kind of a bad cycle of things all around. Some of the same mechanisms that made 23 and 24 brutal for European venture. So hopefully we can find a solution to this. I'd say in parallel, there is upside as well. We've seen just today, The UK government came out with a new report called Building Our Nuclear Nation, and they are really adopting lots of new policies to boost the development of nuclear power, both the small modular reactors and also the bigger power plants.

3:39And the whole idea here is really to rebalance our energy mix away from the over-reliance on gas, make it cheaper again. Today, it's almost five times as expensive to construct nuclear power generation in the UK as it is in other countries like South Korea. And that, of course, is something we need to address. And that's the whole idea of this new policy. So I'd say that's a good thing. And that is part of a wider backdrop you could look at here, which is that sovereignty is becoming investable. That is sovereign energy. It's sovereign technology. It's sovereign defense. The energy transition, which we've been talking about in Europe more than anywhere else, kind of all this greed energy.

4:15But it's no longer just about climate policy. It's about energy security. And hopefully that's something that can mean that we can drive more of the right type of investment. And just to top out the segment, I would say, look, we've seen the EU Industrial Accelerator Act, which has just been adopted, kind of the whole made in EU procurement requirements, which are around making sure that we buy more and we create a better market for buying technologies that are developed here in Europe. So potentially some good things that could come out of this if you squint a little bit. but i must admit the immediate backdrop is tricky yeah i the and the eu by eu they've included the

4:53Lomax Ward:uk in that haven't they so it's not racially we haven't we haven't been excommunicated from the clan which i think is which is a lovely thing lomax obviously we are baked in inertia here this is this is going to be a challenge which aspect would you pick up on a few things to pick up as it as It relates to tech and particularly tech in Europe. So the first thing is, interestingly, the public markets investors seem to have seen that tech is a safe haven. So Mag7 has performed, it was up 1.7 % since the conflict began vis-a-vis other stocks which are generally down. So it's kind of ironic that we've been talking potentially about big tech getting over its skis.

5:33And now people are flying to said people potentially over their skis as a safe haven from global meltdown. So that's an interesting development. We'll see if that continues. The second one is a point that I personally, and I think we've all been saying a lot repeatedly on this podcast, which is around defense tech, which is that Ukraine has become effectively the Silicon Valley for new wave, new age defense tech. So this concept of not developing$5 million missiles and firing them in what you could call old school defense, although we'll talk about that in a minute because certainly the U.S. has been using plenty of Tomahawk missiles in Iran.

6:09This concept of sort of fast fashion aspect of defense, spinning up 10, 20, 30K drones or interceptor drones, using them in either an offensive or defensive way. And Ukraine has obviously built a huge specialism here, sadly, but at least there is, I guess, some upside there. And interestingly, the Americans did not consult the Ukrainians before this conflict, although I think they're now sort of realizing that, you know, with the Iranians firing these$20 ,000,$30 ,000 Shahhead drones and taking out, you know,$300 million radar bases that the U.S. has in the region and obviously airports and critical infrastructure, that the Ukrainians have developed some pretty compelling interceptor drones for the Russian equivalents of those Shahhead drones.

6:55So actually, clearly, it's worth chatting to the Ukrainians, who are the world's biggest expert on this. So I think that's a very interesting development. It also segues a bit just generally into the economics of warfare. You know, the Americans have been firing these expensive 3.6 million a pop Tomahawk missiles into Iran. I think, interestingly, there's a good article in the FT this week where they said that the Americans have only ordered 322 of these Tomahawk missiles in the last five years, where they fired 125 at the Houthis last year, and they've probably fired at least 200 this year. So in a way, they are massively depleting their stocks of these big missiles.

7:31But maybe going forwards, we're going to switch to this more sort of faster, agile, smaller drone-based warfare. And interesting George Bush quote came out in the Benedict Evans newsletter where he says, I'm not going to fire a$2 million missile at a$10 empty tent and hit a camel in the butt. A little bit derogatory, but the point is this, is the economics of warfare do matter and the type and scale of warfare does matter. And clearly these Shahhead drones are causing massive damage locally in the Gulf. Very interestingly, I guess also, but very sadly, in the context of tomahawk weapons. And I think this maybe feeds into the role of AI in warfare, which feeds into the potential anthropic debate with the Pentagon, which is we sadly had two tomahawk missiles hit a school on the first day of the conflict.

8:21in Iran with I think over a thousand, sorry, a hundred children died in that horrendous attack. And interestingly, I read one observer saying on the first day of these attacks is probably when you have the most intelligence at the state at that time. And to make a mistake like that, which clearly was a very targeted attack, supposed to be probably on an Iranian guards, revolutionary guards site, is pretty shocking. And there are kind of, first of all, it's pretty obvious that it was an American missile. There's even like a Raytheon, apparently a kind of circuit board that's been discovered at the school, which is the maker of these Tomahawk missiles.

9:00This is still too early and I don't want to set off rumors or whatever, but talk of like, was AI used in the targeting? And so suddenly, actually the role of AI in warfare starts to get incredibly controversial.

9:11Lomax Ward:You've just thrown a massive brick on the table, haven't you? Right, I know, I know. But Lomax, I mean, just for the sake of the conversation, isn't the counterpoint pretty obvious here? I mean, we've seen from autonomous vehicles that robot drivers are far more accurate than human drivers, reducing the accident rate by 95%. I mean, I think it's entirely likely that AI guided targeting would do the same as so many situations where humans just can't process all the information that goes into things. And so, I mean, isn't it likely that, if anything, the targeting would be more accurate with AI than less?

9:50No, you're totally right. And this is the same with AI for radiology. But, you know, as ever, when you adopt new technologies and you start to have this kind of centaur approach of humans working alongside the AI, like a kind of cobot style in the context of a factory, then in a way you might end up with accountability and responsibility and things falling between the cracks, right? It's like, well, human thinks the AI checked it. The AI kind of did its job. And so you end up, I don't know, who knows? But that is definitely something that we've seen in other industries.

10:22Lomax Ward:Well, and also if the perception is there will be a perception issue, assuming people in the public sphere get to know about these things, there will always be the PR about it. It's the AI. There's certainly a big PR element here. I mean, look, you know, obviously Hegseth has taken a much more aggressive stance than his predecessors. You know, he's re-named the Department of Defense, the Department of War. I was reading an article this week saying that there was a kind of safety team in the Pentagon, which is supposed to avoid collateral damage. 95 % of it's been disbanded. So I do think that there is that whole piece going on.

10:58This is conjecture mainly, so we should be careful. Certainly as AI, because there's both like AI almost in a white collar sense with the way that you use it for targeting. And then there's like the kind of autonomous systems and drones that you actually have or will have more and more in warfare, which are kind of different things and different things to think about. I guess they're all part of the same point. But interestingly, taking back to Europe now in defense, maybe just to finish this. So, yes, the Ukrainians are kind of best in class at this new genre of warfare. And that will only continue.

11:27So I think it's good that that capability is in Europe. But, you know, interestingly, kind of reverse in Europe this week has been the Polish, who obviously on the on the Eastern Front with Russia have just turned down the president of Poland has just turned down one of these big safe loans that the EU had promised. You know, so the safe as a reminder is the security action for Europe, which is one hundred and fifty billion dollars of loans that the EU is going to make at cheap, well, roughly three percent interest to certain countries to bolster their defense capabilities. Right. And Germany have turned this down because they want it.

11:58Germany can obviously borrow more cheaply on the public market. So this is the amount for countries like Poland. In fact, the loan that was going to Poland was 44 billion, which is a third of that whole budget, basically. Why was it turned down? So, you know, in Poland, you have the sort of more kind of hawkish Donald Trump style president, and you have the more pro-EU Donald Tusk, who's the prime minister. So there's an inherent conflict there, which is an internal Polish issue. But the president turned it down because he doesn't want to effectively be seen to be going cap in hand to the EU, having to owe the EU all this money, even though Poland borrows at like 5 % and they would be getting this loan at 3%.

12:34And interestingly, the Polish had only this week approved this new Iron Dome, which has been handed to a Polish-Norwegian contractor for 3.5 billion, which is going to come out of this loan. So that's now going to be kind of kicked into the long grass. The Europeans are still effing about on this kind of stuff when they really need to get a move on. Also, interestingly, Polish Iron Dome is 3.5 billion. the US iron was a golden dome program is one trillion dollar project I don't know I mean I think about like there's a delta and there's like an absolute like an ocean between I'm not quite sure what you're getting for your money there um yeah it's it's gold right it's

13:13Lomax Ward:gold there are so many jokes I don't I don't even know where to start so I won't but I want to stay with Europe and I want to talk about Europe absolutely nailing it this week I think I was quite void and and pleased this week with the amount of good news that came out in the press so we've got uh jan lukun with amy in france raising the largest seed round ever at a billion there's a proposal from the e6 countries that's the eu germany including germany and france to reinvigorate the plans to unify the bloc's capital markets and then asking the eu to govern not the nation so that might be a good thing we've got alan and steve steve sorry that's probably an in joke for me only but alan the health insurance startup founded in 2016 in paris hit nearly 800 million arr raising 100 million at five bill congrats to you revolute getting their for uh getting their uk license after a four-year wait so loads of good things to be thankful for here lomax anything that you would look at or what we should discuss i guess this is europe's first um first instacorn no um we have that team we have the team out of deep mind in in the uk but they haven't actually raised their one billion dollar seed round right so this is actually um uh actually confirmed europe's first instacorn instacorn being an opening round of seed round of one billion dollars i think you know in the us we had thinking machines no that raised two billion and this is in europe so good good on yander coon it's great that it well i guess the uk would say it's not great that it came out of france but you know it's good generally that came out the french generally have an amazing heritage in computer science and math and i guess this is testament testament to that have you seen you know this you know the fields medal right this math prize that gets awarded every four years the french are number two they have 13 winners of the fields medal behind the u.s who won 15 i did not know that that is a very interesting bit of information i know the french are great mathematicians which obviously that lends well to computer science so um janekun thinks that current you know current llms are fundamentally the wrong path to human level intelligence i'm sure mads has a very intelligent thing to say here but instead amy AMI is building systems that, you know, learn how the physical world works from video spatial data interaction.

15:21They plan, reason and maintain persistent memory. So closer to cognition than autocomplete. And they act autonomously in real environments. So not just generate text and images. Yeah. I don't know, Mads, what's your, how's, you know, these are, these are so-called world models. What's your latest thinking as, you know, as our current resident AI bot on, on, on these theoretical world models. Very proud to be, be named an AI bot. Is that sort of next-gen robot?

15:44Lomax Ward:Bod, not bod. Sure, I'll take it. I think it's fabulous. I think it's great. World Models is obviously where it's at. We've often talked here on the show about how we've been constructing intelligence out of words. And words are incredible. Human language is incredible. And we can see what the models can do. And I marvel at what my clock code and my Opus 4.6 does every day, but it only goes so far. And we also know the models can discuss Newton's physics, you know, and explain them at great length and with great poetry, but they couldn't navigate across a room, right? They can't control things in the real world.

16:22They just don't understand how space really works once you get into it. And so you need different kinds of models. And so the whole idea with what Jan McCoon is doing with Jepa, his model at AmiLabs and others are working on is to create models that have this spatial intelligence and innately understand, how the world is actually built up of atoms, not just the words that describe them.

16:46Lomax Ward:Mads, you may have an answer to this. What else is beyond world models? So we've got large language, we've got some obviously small language, other variants, world models. What else would fit into the completion pot? Do you know? I don't know that we know yet, Dan. I think there was a time when people didn't think that LLMs would be more than autocomplete. And I think what they can do today goes far beyond what most people, even many of the inventors thought we'd be able to do with the models. I mean, they are incredible. But it's also clear that they have limitations. And so the idea of a world model, the idea of models with spatial intelligence is the thing that has emerged as the next big frontier.

17:27And it's a frontier. I mean, the map's not been charted yet. We don't know what lies beyond. We don't know whether we're going to go into this world of world models. And that is where the rainbow ends and the pot of gold is, or whether there is another thing beyond it. Of course, we will keep finding new technology. We will keep doing new things. But we've talked, I think we talked at the start of the year about when are we going to get to AGI? And my assertion at the time is there's no way you can get to AGI with an LL. Because until you really understand space, you can't have general intelligence.

18:01But it is plausible that with a world model like what Lacuna is developing, you can get to AGI. You can get to something that understands the world well enough to be able to do anything a human can do without limitation.

18:14Lomax Ward:It's a bad analogy, but it always feels like these frontiers are part of a brain. And we're building the building blocks of a... It is a bad analogy, so I'll just stop there. But it feels like we've got whatever your amygdala and you've got your world models, which is your prefrontal cortex, and maybe some other things. It's not as bad an analogy as you might think. It turns out that if you look biologically at how the brain works and how synapses fire and how different neurons, they work together to think about concepts. Well, actually, LLMs work in similar ways and we can see how they fire and we can see how different regions of an LLM will fire and work together to address specific and complex concepts.

18:56Now, the truth is the biology of human intelligence is still not very well understood, which is why, as I see it, biology is the next big frontier for us. Computer science seems to be largely a solved problem. And it is likely that we can get, certainly software development is a solved problem. And it's likely that we can get within the next few years to something that looks and feels like AGI, a model that is so powerful, it can do anything a human can do. And what is the next frontier then? If we've conquered silicon, if we've conquered the digital world and everything that happens there, well, recursive self-improvement is then limitless, right?

19:31If the model is so smart, it can just get better and better in that world. And the thing that's left for us is, biology. It's the real world. It's understanding how plants and humans and animals and how all that stuff works. So this is incredibly exciting. And so to your point is, I think it's hard for us to say exactly what the analogies are with human brains and human intelligence and digital intelligence, because human intelligence in itself isn't that well understood yet.

19:56Lomax Ward:Yeah, yeah, yeah. Blimey. We're going to get into singularity chat soon, aren't we? Bringing us back to the here and now. I think this is a great, this is a huge thing. this is a huge milestone for Europe, potentially a great, you know, research hub that even if this thing, to be honest, even this thing doesn't really work out, which I hope it will. You have an amazing pool of capital hiring an amazing and fostering an amazing group of talent in Paris, right? And we all know what comes from that. You look at the people spinning out of DeepMind now to create their own company. So I think this can only be a good thing.

20:27Elsewhere in Europe, as you said, Revolu have secured their UK banking license in... Four years. First, well, yeah, in four years.

20:34Lomax Ward:Only took four years. It took four years. I mean, yeah, it's just like classic UK FCA Malays. But great, great for Revolu, which is still on a massive tear. I think still on progress for 100 million customers. Aren't they getting their US license at the same time? Didn't I see that this week? I just don't know what, you know, the latest valuation is 70 billion. I think they're now doing a secondary at 100 billion. I just don't know what the cap is on Revolu, you know? Well, I said we didn't really pick up on London maxing, which is the thing that's been trending here over the last week. I just thought it was beautiful to see.

21:07There's so much happening right now in London's AI scene in terms of investment, in terms of companies. Now, clearly, AmiLabs and Lacoon is in Paris. I think that's wonderful. But if you look at the capital, if you look at the talent that's flowing into London right now, if you look at the scale-up, OpenAIs, Meta is investing here in their AI lab, so much is happening. So I just think that's great to see.

21:27Lomax Ward:Yeah, and there's a number of American R &D outfits that are starting to move back to Europe or move to Europe which I think is kind of interesting. I'm glad to see how that plays out over time. This is all good stuff because it all ties into, you know, these people can create, spit out, create their own companies going forwards over the next five, ten years. So I think that's very, very exciting. A couple of other things I found interesting is, which I think is hugely exciting, is the co-founders of BioNTech, which, to be honest, is the biggest tech company, you know,$25 billion market cap is probably the biggest tech company that's emerged from Europe in the last 10, 15 years, or one of them, Kelly Revolue, 70 billion, but Barntek is certainly up there.

22:09The founders, Ugo Sahin and Oslem Turesi, sorry, there's Turkish names. Well done. I probably thought that. I would have bastardized those. But should, should I say. But they are now announced, Barntek is obviously a listed company, famous for its COVID vaccine, among other things, which is pioneering of this mRNA vaccine technology, which they're now using in the nhs to run trials for cancer vaccines it's a hugely important um company um they have now announced they're going to spin out and create their own startup which you know i don't know if it'll be an instacorn because biotech is not quite why that that as much is not clear but um they have a they haven't well they're running they're running a 25 billion market cap listed company you know that is it the innovation angle is it because they can't move as fast or do as much i'm not going to speak for them but yeah of course if they raise a lot of money in a private company to target something new then you know they'll be able to move potentially faster right so i think that's that's exciting for sure and good to see that he was good at that these guys i had a couple of friends who worked at apple yeah

23:17Lomax Ward:they were good at that they were good they were good at creating the internal competition conflict independence so they could get some good stuff going I had a couple of friends in some of the departments there. At the more macro level, one thing that you flagged in the docket was around the Europe's six largest economies, the E6. Pop quiz, what are the six largest economies in the EU? Come on. I would go Germany, France, Italy. Oh, blimey. What would you go after that? Poland? Yeah, yeah. Poland? Yes. Spain, Netherlands. Yeah, okay. Obviously, Mads knows. So the E6 want to create a central supervisor for stock markets and other critical financial entities.

23:57This is something that's been floated and talked about for a long time as a sort of harmonization of capital markets to make things easier to both raise money and deploy capital. And this was included in the 2024 Mario Draghi report on competitiveness. It's kind of been since then a little bit kicked into the long grass. It's now being picked up again by the E6. At the core of it is a proposal to move supervision of large cross-border stock exchanges to ESMA, the European Securities and Markets Agency based in Paris. So I think this is a, if this happens, it's a really, really good thing. It sounds really boring.