E648 | This Week in European Tech: The Baltics, Bureaucracy & Building Boldly

28:39And something else to say that here is what we should do as a country to make conditions better for entrepreneurs. The truth is I don't think it'll take that much. We don't have to fundamentally reform every single thing. There's some of the big moves we have to get right. You can get UK growth back to 3%. And if you do that, it's going to be an amazing place for founders of all stripes. And there you go. Okay. Okay. Well, listen, we could talk about this one forever, but we are going to move on to a very, very quick topic. UK's quantum opportunity. So Lomax reading this week about two quants.

29:11So quantum and quants, which we're going to come on to next. But just tee up what's happened. Is there something in this quantum opportunity or not? I think this is taken from an article in the FT just talking about the UK playing a key role in quantum technology, particularly quantum computing software and hardware it feels quite a european we've got quite a lot of european smarts in this area and i don't know what is happy is or happy pr but it feels like a lot of that uk aside but across europe there's a lot of good stuff going on with quantum no there definitely is there's a lot of core ip that's been developed um in labs in europe and has now been been spun out and well you know not not to go back to the conversation we just had But if you look at two of the best funded private quantum computing companies, which is Quantinium and CyQuantum, both had roots in the UK.

30:05CyQuantum out of Bristol, Quantinium was a result of a merger involving Cambridge Quantum Computing. And Quantinium, just a very boring fact, Quantinium is top three of the Q3 deals out of the top 10. Quantinium, I think, came in second. But we'll talk more about the success of what's happening in Q3. Yeah, I mean, Quantinium and CyQuantum are valued one at 7 billion or 8 billion, one at 10 billion. And INQ, which is listed, is 23, 24 billion market cap. And all of these companies, by the way, are long way away from any revenue. So just take that in context. Yeah, they're like three years old, right?

30:42They're babies. No, no, they're not three years old. I mean, I think the projects are probably decades old. The companies, yeah, they're more like six, seven years old. Oh, I thought Quantinium was 2021.

30:58Quantinium is a merged entity. But anyway, the UK has two of the leading, aside from hardware, and I'm not going to talk about my own book because both Andrew and I have a very, very interesting quantum company, which is going to be announcing big things very soon. but on software this article is um is covering software and actually two companies facecraft and river lane one again out of bristol one again one out of cambridge um are both becoming key components of the sort of middleware software part of the stack in in quantum still a very very long way to go these companies could flame out etc the article is basically saying on quantum uk not bad and i think that's that's basically right but again we've seen two of the most interesting projects already go to the US and tap those capital markets because they couldn't do it in Europe.

31:46Okay. Well, hopefully we'll see more of that. Quantum is super interesting. Always 10 years off. We'll see what happens with that. Thank you. Another quant. So again, reading this week about quant firms and how traders are crushing it currently. There's a lot of money going through algorithmic traders like XTX and quantitative investors like Cube and Quadrature. Now that each made more than a billion pounds in revenues in the last year. So is this another UK edge? Is it meaningful? Would it save the city, the economy, the UK? Just a little bit of financial context. XTX is spending more than a billion euros building its own data center in Finland, which I found absolutely fascinating.

32:23Now, Mads, you are a resident polymath geek. What's happening in the quants, the other quants? It's a fascinating area. These quant funds are big, big money. There are many implications of that kind of a lot of innovation is happening there it's attracting a lot of really smart people a lot of the investment we're seeing is there and you know one of the things i think we're going to be talking about a little later on i'm just going to lift the lid is this new german ai factory that's being built got lots of money being pulled in and and they're building this kind of cluster with 10 000 gpus to help you know strengthen the german industrial world using AI.

33:04But it turns out that XTX, just for the quant trading, are building a cluster with 25 ,000 GPUs. Okay, so the industrials manufacturing kind of the heart of Europe gets 10 ,000 and the quant traders that just does financial markets arbitrage and makes bucks on that, they get two and a half times as much. It's a good thing in some ways, but in other ways, it is a problem because, you know, many the best and brightest graduates that should be creating technology companies are learned by Jane Street and XTX and others. Brilliant starting salaries, you know, straight out of school, you can get really, really good money going there being a trader, as opposed to going into a startup and building stuff that will change the world.

33:46And that is perhaps a little bit of a shame. Now, I believe in the free market. I think it's a great thing. You know, we've let the city of London down in so many other ways. So this is an important component. But I would say we've got to fix the kind of the talent gap issue. And if all the best quants coming out of Cambridge, they go to be quant traders in the city of London, we got to find other ways to bring the best talent from elsewhere in to build startups because we got to build startups. Nothing is more important than building the next generation technology companies. It's okay. Jona's on it.

34:15She's rocking the Baltic. She's going to bring it all in and solve our talent issue. Although we just said that there's not that many techies as we think there are, which I think is interesting. And then probably we'll end up in like the same situation where you're now, where all the top talent is migrating somewhere instead of staying in Lithuania. So now I'm happy that everyone is staying in the regions. Yeah, the other thing I learned yesterday was that Germany has an exit tax. If you've been there seven years or you're a resident, you have to pay on your unrealized gains before you're allowed to become another nationality, which I thought was a lovely way of kind of keeping people absolutely still.

34:51Norway has the same, actually. You're not segwaying into the upcoming Rachel Reeves budget, budget discussion i promise you i will not do that i promise listeners i will not do that i promise no more no more no more uk politics doom and gloom budget because she is planning an exit tax no so you guys have what have you guys got like you got two weeks left i've actually got some space here i've got a spare room and i've got a surfboard you know i've got i have got a spare surfboard and i've got you know we've got plenty of room in the office guys so i'll put up i'm traveling so much these days i'm not even clearing the 180 day in the uk bar so i'm scott free is that why is that why yeah is that why you're in singapore mads is that is that what you nailed it you nailed 179 now um lomax anything before we move on on the quants now and i obviously know that new york is is supposed to be the center of good um i think this is a good news story for the uk in the sense that you know sentiment would would dictate that the uk is terrible at everything at the moment and we're actually saying well the uk has actually been doing very well on the startup upfront and actually financial services industry in the uk which has been a huge contributor to gdp and jobs has really been suffering in this you know economic disaster period that matt clifford described it as but actually this is one shine of light which is the you know quant trading actually london does remain a leading center for because actually don't forget the last 10 15 years the uk has not only lost out big time to the us but it's also as a financial center lost out hugely to amsterdam paris milan so actually this is a place where uk does retain its crown and actually quant trading has been going on you know for 20 30 40 years and the uk kind of pioneered or had some of the old the old names like winton man group um who started all of this and now handing on to the sort of next handing the baton onto the sort of new wave the xtx markets the cube research and the quadrature, the kind of new wave quant funds.

36:51So that's kind of good to see an evergreen nature to that. Interestingly, if you look at Man Group share price, Man Group, which is the kind of old school OG of quant trading, share price is basically flat for 20 years. Really? So yeah. Quant trading is just part of Man Group's strategies, right? Because they bought AHL. Very true. Very true. They bought AHL, yeah. Yeah. But they also do other stuff. These are big houses. You know, James Street's AUM is 140 billion. which is much bigger than any venture firm. Sequoia is about 80, 90 billion. Maybe GC is maybe 100. Jane Street, we have - And how many times will they lever that?

37:26I mean, they can do some real damage, right? True. We have them to thank for SBF, no? Sam Bankman-Free came out of Jane Street. Maybe he was part of that talent. He's a fantastic angel investor. You know, he wrote the seat check in and profit, didn't he? Yeah. If you've been allowed to keep the shares, you're going to repay everybody so many times over. and interestingly some of this money is being recycled into our industry you know like alex xtx sort of is is putting a lot of money into yeah we just probably don't deal with it xtx is probably one of the most prolific ai investors in in london at the moment so i'm at the seed stage so good it's coming back down and actually mads yeah it is taking out the best the best computer science and maths grads but you know what it's taking them a lot of them away from banking so that's not such a bad thing.

38:13So there is every cloud. Well, we'll see. I wonder when quantum comes online, what happens with quant and with more AI developments, what's happened between them all? So that'll be an interesting one. Talking of AI, Mads Deed, it's AI Corner. So we've got loads in AI Corner this week. We've got more circular funding from NVIDIA. We've got this Calm model from China, this continuous auto-aggressive language model, which is potentially a paradigm shift, but I'm going to let you shoot that one in the face. OpenAI stopping legal advice. Is that a new story? I'm not so sure. We have NVIDIA and Deutsche Telekom building that 1.2 billion euro data center, which you've already touched on, Mads.

38:54And Nebius launches Token Factory, which is an aggregation and infrastructure for open source AI models. So Mads, all of that, go, 10 seconds, boom. We've sort of started opening the topic a little bit of AI. It's hard not to touch it. OpenAI, as you said, did a deal with AWS and Amazon,$37.8 billion, seven-year cloud deal. What's a billion between friends? Let's take 38 of them. Amazon shares really liked it, popped more than$100 billion on the announcement. There's a lot of continuous talk about this Circulus Funding question. How is Sam going to fund his hundreds of billions of promised spending with absolutely no profit whatsoever?

39:35He's lost you know, many, many billions this year alone. And he said, you know what, whether I burn 500 million, 5 billion, 50 billion a year, I don't care. I'm going to solve AGI. I'm going to solve all the problems of the world. So I think there isn't that much new to see here. Other than that, that he very specifically is not letting himself be locked into any one company. We saw how he took a big investment from NVIDIA and turned right around and then decided to spend some of that money with AMD. So, you know, kudos to him. He is absolutely going all in. And I think it's just a really exciting entrepreneur to watch.

40:13The Calm model that you talked about, look, there are new papers coming out all the time. I think this one particularly, I've got to spread out by now. I mean, it's not why it's decided. It has limited interaction on Hocking Face. I think it is one more potential technical improvement, but there are so many, there is so much incredible stuff coming out of especially China right now. But not a deep seek moment, Mads? You don't think it's a deep seek moment? No, no, no, no, no. I don't see this as that. There's another one. There's actually quite a lot of clickbait here because there was sort of this whole open AI is no longer going to provide legal advice.

40:49I saw there were people discussing. I think that's pure clickbait. I mean, they were just, from what I could see, consolidating existing policies to say, look, we're not providing legal advice. Make sure you have a real lawyer if you need legal advice. But in terms of how the model works. You can still go and make it produce draft documents. You can analyze contracts. You can do all that good stuff. That actually is really useful, not to replace a lawyer, but sometimes you just need a quick take on something or you don't have time to get a hold of a lawyer Sunday afternoon or whatever. You need something analyzed quickly and it's better than nothing, right?

41:20And in many of those situations, I think it's really useful. And it potentially opens a new business model for them at some point, like lead gen for lawyers. I mean, because actually they'll have very, very... I think so. Yeah. Yeah, I think so. There's a marketplace in there somewhere. Well, that's the direction of travel for the business. While Anthropic really has captured the corporate market, the developer market. Anthropic has more market share in developers and API use now than OpenAI does. But OpenAI is a consumer product and it's a replacement for search. And it's a challenger to Google.

41:56And boom, there you go. Lead gen, ads, anything that points to transactions and commerce. I think you're right. I don't know if you guys saw it, by the way. There's just been a huge fracas between perplexity and Amazon. I didn't see that. Yeah, perplexity. They've got the whole shop with perplexity. And Amazon is basically like, if you want to shop with Amazon, you've got to be on our side and get our experience. You know, don't let perplexity disaggregate us. Because, of course, what's the first thing that happens then? Power travels somewhere else. And Amazon can be swapped out for other alternatives.

42:31So they're basically like, it's not in terms of service. You're not allowed to use this if you're a droid. So you go to perplexity and you can buy products through Amazon Store? You can tell it to shop for you. Nobody can tell it to shop for you. I need X, Y, and Z. Go buy it for me. Pay for it. Get it shipped here. As in find the lowest, find the quickest, find the greenest, find whatever your sort and filter is. Do it on those bases. Okay. And of course, you can say as a consumer, it should be my right to do this. I can sort of also see why Amazon is doing what they're doing. They've got such a powerful position.

43:06They're the default for almost, at least all men shopping. I know women are a little smarter. They like to sort of look at different places. But I just want the, you know, single click shopping, right? That's how I buy 95 % of the stuff I buy. And the last thing Amazon want is to have that disaggregated by somebody else. A disintermediated, I think is the right term. German data centers, NVIDIA and Deutsche Telekom. They're launching a 1 billion euro industrial AI cloud. Exactly. Now, once upon a time, a billion euros, dollars, sterling, sounded like a lot of money. Today, it's pocket change. We're talking about 10 ,000 GPUs.

43:46And part of me is like, great. It's better than nothing. But also just feels like not very much in the grand scheme of things. And the scale comparison kind of exposes, I think, the ambition gap a little bit here. When you compare that to the hundreds of thousands of GPUs we're seeing installed in the U.S. data centers. I wanted to know if the Nebius token factory was a thing or not. Is this going to be a game-changing thing or not? I didn't know if you had any thoughts on that. Nebius is actually a really cool company. So after the Russia invaded Ukraine, Yandex spun out this needless cloud and it became kind of an EU neo cloud.

44:33So Amsterdam, the Dutch incorporated in the Netherlands, they've captured the balance sheet. They've signed this big deal now with Microsoft. It'll be somewhere between 17 and 20 billion dollars over five years. So that's a meaningful kind of revenue for a neocloud maker. And in some ways, they are the European parallel to CoreWeave. So it's pretty cool. They've got data centers in Europe and the US. It's a profitable business on a, well, okay, not quite community adjusted, but adjusted EBITDA basis. I don't know whether that's because they're running 10-year depreciation schemes or what it is.

45:12Maybe some accounting artistic license. But the numbers are better than CoreWeave's. And now there's a lot of revenue coming too from Microsoft. Challenge perhaps is that with the Microsoft revenue, it looks like they'll have as much as 80 % to 85 % revenue from one customer. So let's hope they can get some more in. A bit like CoreWeave. CoreWeave is slightly less. It's between 60 % and 70 % from Microsoft. And then OpenAI is 15%. isn't nebius the biggest like one of the most important tech companies in europe that no one's ever heard of no i think so i think it feels like it which is why i wanted to ask the question it feels like you know this isn't something that you read about in the press and is that hang on a second this is like in fact tell us tell us more about what it is and why it's important well nothing i i i think that um well as as mad said it's a neocloud company i one thing that people are seeing Nibis as is a answer to European sovereignty when it comes to cloud and compute, which it may well be because it's a European headquartered company listed in the US.

46:21The irony is that the answer to Europeans tech sovereignty problems comes from a forced spin out from Russia's most successful tech company. It's not lost on me or any of you by the looks of things so um but i think the open source models chinese originally the models which models isn't this isn't this an aggregation of like 60 different os is the infrastructure in the model don't so you what you're touching is something we didn't quite get to unpack which is this new thing they're launching this token factory but they're basically saying we're taking all this panoply of wonderful chinese open source models we're putting them in the cloud you can get incredible tokens at really attractive prices.

47:05It's going to be much cheaper than Anthropic. Almost as good, if not as good. Let's go. And I think it's a super nice story. It's not, this is kind of what we've been talking about the last year, right? Which is that Anthropic has been right on the bleeding edge of what you can deliver, but it's incredibly expensive. And if you are any of the application layer companies like a lovable and you're spending all your revenue with Anthropic, well, what are you going to be looking for? you're going to be looking for a displacement or kind of a placement model and a token factory could be just a ticket.

47:38So I think the industry here is evolving like we thought it would. Question is, how will the frontier labs respond? They've got to respond in some way. Somebody comes in office, you use something that's 99.5 % the quality of what you do at 10 % of the price, it's going to be an issue. So it's interesting to see how that develops. in context, if it's good enough at 95.5 % or whatever, that's all you need. And if you have access to this super low cost, and these are all managed, hosted in incredible infrastructure and data centers, that feels like a really compelling proposition. So that's why I found this story interesting.

48:22And I think we may not see it overnight because the truth is, if you're a developer, phone, okay, and you're spending 200 bucks a month on cloud code. Do you know what 200 bucks a month for the productivity gains you get? It's fine. It really is. Like whether you spend 200 bucks or 100 bucks a month, nothing to see here. Like the productivity gains are worth tens of thousands of dollars. But lots of workloads are going to sit beyond that. And there's going to, people are going to start finding more and more places where that 99 % as good model will be, you know, fine if it's a 10th of the price.

48:59And so you will see workload displacement over time. And of course, this is the thing we've been talking about for some time. Okay. Okay. Lomax, Yona, any more on the AI piece? Anything grab your eye? Anything else that you'd like to dig into, Yona, starting with you? No, I would just probably like finish on the Nibius, which is obviously if it's a from Yandex, it's like very clear, like in our region, it's a Russian company. You know, like this is kind of the mentality, how we view it. But very honestly, I haven't heard of it like much. So I think they have done that perfectly. It's a bit like the car model hasn't really hit the press, maybe for the right reasons, Mads, before you shoot that one down.

49:42And it hasn't really. My take is because the foundational models from the States capture all of the oxygen, we don't seem to pay attention to maybe some of the other fascinating stories companies and models and technologies that are there around there which is part of the point of this pod just to just to give it a little bit of a bit of a prompt one general observation is i think there's a very lazy piece of analysis going around which reminds me of a sort of old man in a pub sitting there in the afternoon, which is around this circular financing, is that whenever any of these deals get done, everyone just goes, oh, circular financing, we're in a bubble.

50:26I don't think anyone is looking actually at what the substance of these things is. And I don't think we're going to get into this debate here, but if OpenAI is booking ahead some compute with AWS, that's not circular. So as in like, just when one AI company does a deal with a hyperscaler or with another AI company or a chip company, it doesn't mean it's a circular arrangement. If you go back to the classic definition of a circular arrangement, or you look at the vendor financing that NVIDIA has been doing, which General Motors used to do with consumers to buy cars. They used to lend money to people to buy cars.

51:05That was a very accepted thing. So I'm sure there may be, if you spend a lot of time looking at this, some things that raise some question marks, but I just think there's a sort of a lazy piece of analysis that flies around everywhere. And you see it a lot on the social media, writing any of these deals off a circular, when actually I don't think it's as, the analysis is as simple as that. It feels like it's a much broader observation rather than looking at an individual flow. The web, the intricate web of who is buying who is investing the flow of money i think there's my take is that it's just an interesting narrative but yona sorry i interrupted you please carry on no no all good but i completely agree here with lomax i think there hasn't been a podcast last week that we didn't show a picture of like uh the bloomberg plugging itself yeah and it was like all over the place.

52:00It's nice to debate, but I agree here. This might age very badly when the whole house of cards comes down. Just for the sake of argument here, because arguing is fun. There's a good old rule of thumb, which says that at some point cash needs to change hands. Otherwise, it's not real business. And I think one of the challenges we have here is you've got Microsoft offering open AI free credits as investment. No cash is changing hands. Open AI then turns around and uses the credits at Microsoft. No cash changes hands. And so I think those are some of the things that look a little bit iffy, right?

52:45If you think about it. So Microsoft is turning their cash into equity in open AI without ever sending any cash in a round trip, which means you're disabling some of the normal free market economy a little bit. You're sort of, you're creating some weird structures. And I can see why people would ask questions of that. But I think there's some things that are completely different, which is NVIDIA, for example, saying, listen, we've got so much cash on the balance sheet. There are all these amazing companies out there. We want to put some cash into them and invest, right? We want to buy a slice of OpenAI, and OpenAI turns around and spends some of the money with AMD.

53:21I can't see any problem there. I mean, other than it's just, Gen's means like it's a generational company. I want a slice of it. What's wrong with that? Yeah. Well, let's talk about generational company. So let me just set the scene with deal of the week. I was reading this week about how Europe has broken all records across AI defense and quantum in Q3, as in investing. Deal making spiked, seven new unicorns, two jumbo AI raises, the top three deals being Mistral, N-Scale, and Continuum that we were talking about before. Now, between just those three, there's been four billion raise. That's two UK startups and one French.

53:58So that's the scene. It's been a bit of a bumper Q3. I think that's super cool. But, Jona, what was your deal of the week? It's probably not the one in billions. That's okay. We can take millions. We can take millions. Yeah. And when I was thinking of the deal of the week, it's actually like more of a deal of the past week, which was announced and probably done like back a few months ago. but it's nexus yeah yeah they all are that's actually true but it's nexus ai so it's the founders of tessanet which includes nordvpn and many other companies that i've mentioned that are bootstrapped they've uh they're building like uh ai uh different ai llms rerouting for for the companies which actually there are quite a few other companies doing that but they basically raised on the pitch deck they raised eight million and then later from index ventures and then very quickly were followed up by eventic capital creando and divantic capital is actually one led by matt miller ex-partner of sequoia uh so they've had a challenge this week haven't they as sequoia saw as all kind of had a bit of a reshuffling at the helm oh yeah yeah i'm not surprised you know that's kind of the legacy house.

55:16So yeah. Roloff Bonsai, he says one week that Venture Capital is done in the asset class. The next week he's out. Right? Double teacher. Okay. I'm almost saying that so I should be careful about that. You won't complain if they mark up your deals, Giannis. Exactly. No, no. I love it. It's really nice. We have Sequoia Scouts all over the place so that's good. So yeah, so that's I think like the deal and not that they need money. They're printing money like crazy and they just want network. They have networked themselves as well, but they want to, you know, be there on TechCrunch, show competition that they are there.

55:57It's really great. And, yeah, I think it's really good news for the region. Did you try and get into the deal, Yoni, or not? No, too late. Yeah, no, we did not even try. Like, we asked, do you invite local fans? And they said, like, no, it's basically. At least they took your call. They could have just ghosted you. Actually, yeah, it's true. But this is what I said about the network. We are all so close here that it's very easy. And I wrote on Signal, I asked, they said, no. Lomax, what's on your docket? This is definitely the billions. So NVIDIA is preparing to invest up to$1 billion in poolside, a sort of French-American company.

56:41his valuation is set to leap to 12 billion, a fourfold jump from its 3 billion valuation last year. And this would form a big part of the 2 billion fundraising round, which has already drawn more than 1 billion. So what the 2 billion round, 1 billion from Nvidia. Poolside is kind of an anti, I guess, anti-vibe coding. Like it's really focusing much deeper on the automation and deployment of code. Founded in 2023, it's a very young company as well, founded by Jason Warner, the former cto of github and iso cant very impressive um company well let's see how things go but that in the context of europe is a is a mega raise you know 2 billion a 12 billion valuation um is one of the biggest rounds in europe this year so very nice to those guys and what i would say if you want to if you want to kind of a bit of behind the scenes i got to know iso uh a few years ago because he he spent a lot of time in lisbon but clearly i didn't spend enough time with him because I didn't get access to pool sides.

57:41What the hell? No surfing for you two. What the hell? I know. Don't order Lomax. I spent enough time with Nexus founders. But congrats to those guys. And they actually do, because ISO does have a bit of a link to Lisbon, they have a big role in the AI hub here, actually. So I think they have a bit of their team here as well. So definitely a nice story, which kind of, going back to the conversation about entrepreneurs first, you know, building big global tech companies. There's got to be a transatlantic element to them at some point. And, you know, Poolside is a really true transatlantic company.

58:17Well, lovely people. What is happening this week before we disappear? So I'm off to Turkey meeting some founders and some LPs. So this is my first trip to Turkey. I've not been before. So I'm looking forward to that. Yona, what are you up to this week? I'm always with home. It's kids holiday. They want to go to countryside. Oh, we've just done that over here. We've just had all that. Yeah, and then we have some LP dinners planned because we are fundraising. So that's a little bit on the work side. But yeah, no turkey for me, unfortunately. I've heard it's great now. Yeah, I'll feed back next time we speak.

58:55Lomax, what are you up to? Well, Lisbon's got Web Summit next week, which is the big tech conference. Although I do feel that Web Summit, you know in the context of startups and and gps and lps early stages it's not now a kind of big tech conference for yeah and it's obviously now no longer cancelled right because there was that big fracas the ceo said something but it's never it's never recovered from that i remember web summit in 2022 all the big us funds were here everyone and to be honest if you look at the you know you look at the app on early next week and the gps and lps coming it's um it's slim pickings to be honest it's never recovered from really interesting everyone is going to slush now yeah it's the following so slush is the is the week after i think interestingly i i think in a in some kind of in i would call probably fantasy marketing i got an email from web summit yesterday saying oh if you're coming by private jet to web summit um unfortunately there are no more slots left i mean i don't know anyone coming to web summit on a private jet i suspect there is no one i I suspect it's a bid from the organizers to appear relevant still.

1:00:07Yeah. Mads, what are you up to this week? Next week, speaking at a conference in the Middle East. So back on the road. Jesus, Mads. Yeah. He is avoiding this exit charge. Yeah, yeah. The 180-day king. Guys and girls, thank you so much. Yona, thank you so much for joining us. Gentlemen, always a pleasure, always a gift. And we will catch you next week. see you in the next one bye bye thank you