E433 | Jonatan Luther-Bergquist, Inflection: The development of crypto to sovereign compute

0:00We're no longer in this unipolar US versus maybe lesser enemy China or Russia. it's now completely multipolar meaning there's no clear dominator in the world this can be extremely dangerous especially for europe which is not exactly the the winner in this we haven't caught up yet we're still catching up i would say the way to do this is by making sure that we get access to the type of compute that shores up our our sovereignty um that's the basically the only leverage that we have like it's very asymmetric we need to get better at three things and that's scale of compute like making sure that we can do more compute in every aspect of the world so more physical semiconductors but also do better compute and achieve more things with the computer that we have we need to get resilient resilience of compute also meaning we need to protect our physical infrastructure using compute in the best way also against external manipulation also understanding what AI does, how it affects other compute systems.

1:07And then finally, we need to equip all these compute resources, the machines, with high-quality data and sensors. And we need to guard that. We need to establish some form of economic guardrails that let machines transact also and share those resources across various types of heterogeneous compute resources. This is very much going in the direction of ultimately integrating everything that we do, everyone, every aspect of our daily life with computers. And only if we really control this are we truly sovereign. Here's a few words from our beloved sponsor. If you're looking to meet key players in Central Eastern Europe's private markets, this is the conference you don't want to miss.

1:49Join us for the 14th edition of the premier networking event for Emerging Europe. You'll meet investors from firms like Antenna Digital Ventures, EBRD, EIB, Earth's Group, IFC, Tybus Ventures and many more. Join 0100 Emerging Europe from May the 15th to the 16th in Budapest.

2:12Tear down this wall. It's more than just an alliance. This is a union of values. Let's start acting. This show is not investment advice and the hosts of this episode may be invested in the funds and companies featured. Welcome back everyone to the European Easy Podcast. Today I'm joined by Jonathan from Inflection. Jonathan, welcome to the pod. Thank you so much. I'm very happy to be here. So, Jonathan, Inflection, I know you. I've been an avid follow-up of your writing as well as also your co-GP's writing, but not everyone necessarily does. Maybe you would start us by telling the story about inflection.

2:58Yeah, absolutely. So to be fair, I wasn't there from the very start, but I've been here for the majority of the investments that we've done. So let me tell you maybe shortly where we're coming from and what the background is to the name also. So does that have a significant meaning to how we invest and how we think about the world? Inflection started out as my co-GP, Alex Lang, who is a lawyer and economist by training. He was working in venture for a couple of years and was always focused on sort of the frontiers and the fringes of technology, the next larger thing that is coming. Like he wasn't able to tap into those in the way that he wanted to.

3:42and so decided to leave. He was at Index Ventures last before he started inflection to leave Index and go out and raise his own fund. That was in 2019. So together with our partners in the back office, he raised the first vehicle. That was back then the most, I'd say, interesting novel technology that was really changing the fabric of society, of digital society mostly, was really around crypto. So the first couple of investments were very crypto heavy. We built a strong network in that space. I would say there was a massive inflow of really high quality talent into that space as well. If you didn't want to follow the traditional path of starting a playbook or something that has a napkin attached to it, that was kind of what you did because that was a very weird, very strange thing to do.

4:37Start a crypto company in 2019. that's where we saw the real venture side outcomes. So what we think about venture is it has a place as a product because it has this extremely convex profile. So what that means is we can make a rather small bet at the very early stage and have an extremely large outcome. It doesn't happen very often, but it does happen if you put it in the right space and in the right type of company. Back then, that was very much crypto in 2019 in our eyes. To add some context to that, that's exactly the question I asked you when we had our first chat, right? So I know inflection for being all about crypto, but now you're very much about sovereign compute, which obviously, and we'll talk more about that technology-wise and trend-wise, looks like crypto in many ways or has parts of the crypto, both community and technology stack, that is similar.

5:42However, it's also considerably not similar. So my question to you was, tell me, walk me through how you moved from being known as a crypto fund to now being about sovereign compute. Yeah, so this has two aspects. One is the marketing aspect of it. How are we seen externally and how are we branded and so on? And the other aspect is, how do we think about it internally? How do we reason about what technology achieves and what it does to society and what our role as a rather boutique venture fund is in that space? Maybe tackling the internal piece first. And I have to go back even further to 2016 when I started out in crypto.

6:28That was my first job, actually, after graduating as an engineer, was I got to use what I had done in my masterpieces, which was privacy on blockchain. And I got to apply that to a context that had massive impact. I couldn't identify any other opportunity in the space of all the jobs I could have done or companies I could have started that could achieve the same thing. So I chose to work at a very small company here in Munich. I was the first technical hire. I was implementing the first prototype of a product that then became a financial backbone for the World Food Program. We delivered through this product in the end over three years, we delivered millions and millions of dollars of aid to refugees in Jordan.

7:12My main motivation for doing anything in crypto at that time was, wow, we have this amazing new technology. It can do these 100x improvements over existing bureaucracies and digital products. This is exactly what I want to spend my time on. That meant distancing myself from the sort of Bitcoin speculative, back then it was Bitcoin, like speculative coin ICO bubble, you know, financially. Maybe that was stupid of me, but from a principal's perspective, I think I couldn't have done anything different. And that applies today very much. And then the investments that we've done, we want to help create long lasting companies that have an impact.

7:53I'm not using impact as a sort of greenwashing word or anything, right? I want them to produce products that people use and that have a sustainable business model. The crypto investments that we did all had a very strong computational component, right? Those are main motivation and view on crypto as a world computer. We saw new type of cryptography and distributed systems theory being applied in the real world and really building the new internet. That's very much how we still see crypto today, although that part of crypto is a very minor part of all the crypto projects out there. So the speculative DeFi applications, they fit a very different type of investor, I think.

8:42And that's not us. We look still very much at cryptography and the enabled businesses from having access to these new types of trust models. And the same thing goes for other types of non-blockchain related applications that we look at or companies that we look at. We look for these inflection points, really, in terms of technology, society, founders. just at the beginning of this takeoff, really? What is the thing that has to change for this idea that probably existed for a long time to become reality? To make a concrete example, we invested into a stratospheric airplane company about two years ago.

9:30Stratospheric airplanes have been around for, people have tried it since the 70s. And it didn't work for a number of reasons, right? Like Airbus did one. I think Boeing tried to do one. a couple of smaller initiatives tried to build some, and they just crashed all the time. It was super expensive, and they crashed. And then this new company comes out of Amazon Prime Air and says, we think we're at the right time now, and we have the right idea to build one that works. We studied all the ones that didn't work, and now we kind of know why they didn't. The inflection points were battery density and cycle stability.

10:07so I can charge a battery many, many times, meaning I can stay up for a long time and recharge during the day. And the computational modeling of building such an airplane is possible today because we have the resources. We can understand what forces the winds and the weather up at 20 kilometers altitude due to an airframe. We couldn't do that 10 years ago. And we also have massively better and cheaper solar panels, right? So we can harness that energy. This is really what we look for in every investment that we do. Do you know, just because you talk about inflection points here, and I've often thought this when I saw your name as well, do you know Mike Maple's framework of inflection point theory and so on?

10:56I've heard of it. I know there are similar. I think we think about it similarly. Yeah, I can hear much of the same nomenclature and also how you're thinking about how the different things, both societally, technologically, and also in the founder's journey, match up to be three powerful inflections that match up and then create an opportunity that you as a VC want to back. Quite interesting, I find. Then let me just ask you, because I want to go a little bit back to crypto, just because it's great to talk about crypto with someone who is maybe not knee deep and highly incentivized to still prop up the industry.

11:40Because what you described is there's parts that I really love and I think are going to be incredibly important and also within your thesis. And then there are parts where you've kind of distanced yourself and said, no, this is not us. And I'm thinking maybe you'd be up for sharing a bit of the dynamics in this part of crypto that you're saying, I think it's unsustainable. I see great founders turning away from it. I see ourselves also not wanting to play that game. Maybe starting with a founder perspective, because that's the most important, in my opinion. We have misaligned structures in general.

12:18There are pushes from four very clever people, or there were pushes from very clever people, people with PhDs in math and physics to go into crypto. It appeared like the new shiny thing that you should be working on if you're kind of a successful person. And it also seemed like there was abundant money, which there was still in terms of venture funding. That is great, but that doesn't match up with the reality a few years down the line. If we look at many projects that were sold or marketed as impactful, long-termist projects that will change the world or change the internet at least, a lot of them end up launching a token and then that at some point goes to zero.

13:08Then the question is, what's the incentive for the founders to stick around? In the normal venture world, we have equity and that's the main incentive. And that's aligned over 10 years until you IPO or get acquired. In crypto land, there is a high push from or a very strong push from VCs and investors to launch a token as soon as possible. Because the sooner you can extract the value, the higher your IRR will be. And then you can go on to the next thing. That does not match up with creating a sustainable business model, something that actually works for a long time, that is incentive aligned with creating great marketing machine.

13:45That's not what we want to invest into. Others, more nihilist investors are free to do so, but that's not how we want to... The types of businesses that we want to have on our portfolio website, if you look at them, I don't want to see a launch token and token price down 99 % on all the projects in five years. Some specific stories are founders who tried to raise the next round, have amazing traction, but are met with the somewhat jaded larger funds, specifically in the US or in Southeast Asia. and are being told, we will back you, we'll happily invest into you if you promise that you launch a token and that we'll get X percent of that allocation.

14:38And if the founders don't want to sign those docs and say, actually, we don't know if that makes sense for us as a company or if we want to be attached to that type of company, then it becomes very difficult. And that's an issue, right? Maybe, Jonathan, just to those that are a little less well-versed in this, could you explain the dynamics there? Why is it that the launch of a token would be one of the requirements that's nestled in the terms for a funding round to happen? Yeah, so it's not always a bad thing. There are tokens that have value and fulfill a function. And I mean, look at the whole stablecoin category of companies.

15:23This is one of the examples of amazing technology, just supremely superior to any of the existing infrastructure that we have. But aside from that, if I am at a seed stage company, I might not know exactly how to build my customer base yet. I may have some interesting technology, but I don't know what the incentive mechanisms should be. So this goes pretty deep into game theory and to not to go too deep, but it's essentially I program a rule and that rule should make people do certain things. So incentivize people to take a certain action by rewarding them with some additional percentage points of a token.

16:06And the VC's case, specifically in crypto, the incentive is I want to make as big of a return as soon as possible. So if you look at some of these crypto companies, think, okay, what could be the outcome? What's a potential exit path, say, after a Series A? and they say, actually, the only thing that we've seen has been one or two IPOs, Coinbase, for example, a few acquisitions of medium to small outcomes. And then we've seen all of our returns basically driven by token launches. And after I've sold off my tokens, I don't care at all what happens to it. Definitely what we've seen in real life, even in some cases, blatant kind of nihilism of we want you to launch a token as soon as possible.

16:53We want to get as big of a share as possible. And we don't want to have any investing involved, which would prolong the time period during which you can sell. I think everyone that's listening in here will understand exactly what you're saying and be able to fill the gaps. I think it's just so important that we have this honest conversation in a place like this about what's actually going on in crypto land when it comes to this type of crypto, this part of the crypto landscape. Let's, Jonathan, go to the sovereign compute thesis that is then the overarching name or the category that you work under now.

17:32Tell me everything about it. What does sovereign compute mean? Yeah, absolutely. So we have seen in the past decades, actually, a massive decrease in productivity, specifically in Europe, but also to some degree in the United States compared to China. Now, not becoming too geopolitical, but we realized we need a techno-industrial renaissance. That's the big headline. Now, how do we achieve this? How do we make ourselves more productive? How can we shore up our digital world and our physical world at the same time? There's across the entire spectrum of technologies. I think there's only one thing that ties everything together, and that's compute.

18:23Computers using mathematics to describe and act on the world. They change everything about how we think, learn, build, how we fight even. They already reshape our brains. I have to think about, of course, some negative aspects, but how many phone numbers do you remember today compared to 20 years ago? My own? Yeah, exactly. Yeah, I don't even know my wife's phone number, right? She hates me for it, though. Yeah, sometimes I think it would be practical. Maybe I'll just get a tattoo. Yeah, yeah. I think so, too. It reshapes our labs, our factories, our energy grids, our battlefields, all of our institutions.

19:06So if we look at the Industrial Revolution, where we all of a sudden had access to energy through steam power, that was a huge change. If we look at the early innings of computers in World War II and onwards towards the one in Vietnam, compute started to have an effect on this. started to affect not just the actual battlefield, but how these companies make progress. If I could put more semiconductors in the hands of my researchers, they could be more productive. They could discover new things that would then change the balance of power. Onto the 1990s, we had this clear hegemony of the US with them dominating everything from the internet, mobile, onto socials in like 2010, so on, building the cloud, essentially, everyone else playing catch up with the US.

20:06At some point, China caught up, right? Now, we're no longer in this unipolar US versus maybe lesser enemy China or Russia. It's now completely multipolar, meaning there's no clear dominator in the world. This can be extremely dangerous, especially for Europe, which is not exactly the winner in this. We haven't caught up yet. We're still catching up, I would say. The way to do this is by making sure that we get access to the type of compute that shores up our sovereignty. That's basically the only leverage that we have, it's very asymmetric. We need to get better at three things, and that's scale of compute, making sure that we can do more compute in every aspect of the world.

21:01So more physical semiconductors, but also do better compute and achieve more things with the compute that we have. We need to get resilient and have resilience of compute. Also meaning we need to protect our physical infrastructure using compute in the best way, also against external manipulation, also understanding what AI does, how it affects other compute systems. And then finally, we need to equip all these compute resources, the machines, with high-quality data and sensors. And we need to guard that. We need to establish some form of economic guardrails that let machines transact also and share those resources across various types of heterogeneous compute resources.

21:44This is very much going in the direction of ultimately integrating everything that we do, everyone, every aspect of our daily life with compute. And only if we really control this are we truly sovereign. Meaning on an individual level. Three parts, scale of compute, resilience of compute and integration of compute. Or how do you call it? We call it flow. So flow of data. Interesting. Okay, let's unpack each. First of all, I want to ask you a question. Whenever I think about verticals for VC firms, I try and think about why does it make sense to cut this out as one specific vertical in the zone?

22:34B2B SaaS, same playbook, same customers, same kind of ecosystem of founders, same business model, same go-to-market generally. for that reason, made a ton of sense to have a B2B SaaS-centric VC. Because once you've seen, I was about to say, once you've seen one, you've seen them all. That's not true. But you definitely build up repetitions within that space very quickly where everything fits in together nicely. Sometimes I don't like to hear too broad deep tech VCs talk because I kind of think that, Well, the only thing that's the common denominator in deep tech is it oftentimes involves hardware.

23:14That doesn't necessarily, but it's super hard. And there's more of a tech risk than maybe a market risk. But that doesn't really tell me much about how you as a VC will win. The researchers come from different places. The go-to markets are different. The technology is completely different. So, for that reason, like the fact that you focus on deep tech doesn't tell me very much and doesn't make me very comfortable believing that a two-person team or five-person team or whatever will get access to the best opportunities and will be able to be super value-add here. Why does it make sense to cut out this thesis of its own?

23:51Sometimes it's for LP reasons. But why does it make sense in your mind to cut out the thesis of sovereign compute? Yeah, we tried to not let LP meaning or LP... Yeah, that's the right answer. It does not always the true answer, but I think it's true for you. Yeah, exactly. I think we tried... Sorry to zoom out again, but I have to mention this because it's very, like, at my core and our core as a fund, right? The first fund was 3x oversubscribed. We could have done a much bigger fund. But we said that doesn't fit with our strategy. There's a reason why we're$40 million and why we invest where we are.

24:35There was one very impactful book for me as a person that I think I look for these concepts in everything that I do. I don't know if you read Zen and the Art of Motorcycle Maintenance by Robert Persig. I have not I have heard about it actually because it has a weird name compared to you think it's just a small handbook type thing or it does not sound as it's something a big philosophical work but it actually is exactly yeah it's an amazing journey of a book I would say and I highly recommend it to anyone whether or not you're interested in motorcycles or Zen or philosophy I think it's very well written.

25:21Zen and the Art of Motorcycle Maintenance by Robert Persig. P-I-R-S-I-G. We'll put a link to it on u.bc as well. Perfect. So there's this concept of quality. It's literally about motorcycle maintenance. Like how do I repair my motorcycle appropriately? Because it breaks down and he has to do that in the story. and then goes often in this very long tirade about quality and what that means. So he defines it in two different types of qualities, dynamic and static. So dynamic quality is kind of instantly recognizable, beauty, excellence. It's like, oh, this has value immediately. It's obvious to everyone who sees it.

26:06And then there's a different type of quality, which is called static, which is a deeper pattern, something that isn't visible at the first view, but if you go a few steps deeper and try to understand them, you realize that there is quality. Yeah, like with a big Q. And specifically, you know, this can be social or cultural, like we form laws or we form cultural patterns or groups, but also organic, inorganic, so biological cells organizing themselves. Or, you know, I would say even how we build semiconductors has this quality. It's this beautiful pattern of excellence, really. How can we as humanity go from living in caves to literally taking sand or a compound of sand and make these nanoscale structures and then use it for something?

27:03And there's also the intellectual quality, which now going back to us as a fund, what we look for is this type of quality in each founder, in each idea that we evaluate. If we don't see that, then that's very clear, but it does take preparation to see it. Specifically, the static type of quality, we can't assume to always find the dynamic quality. Those are kind of the obvious ones. They're getting picked up by other funds. We can do the other thing. We can go a few levels deeper by preparing properly and having the context. And that's why I think Sovereign Compute, now again, going back to your question, this is where we as a fund have a reason to win or experience at least, quote-unquote, win the hearts of our founders.

27:51we have the same frame of reference as they do if they fit our fund right there are founders that we just don't fit with right and that's totally fine there are other options for them usually but there are certain founders that when we explain our thesis to them which is always part of our kind of first intro call which are to tell them who we are and so on they say wow yeah i never heard it explained like this before but it makes total sense i think we're very aligned this may sound like a very soft criteria, but a lot of what we do is building rapport in the early stages, trying to understand how founders reason, how they think, and ultimately how they evaluate ideas.

28:31If they think our ideas are great, then they like us, perfect. Now we try to evaluate their ideas, and for that we need to do deep research ahead of time and to be prepared when coming into a call so that we understand what is this intellectual quality? What is this deeper pattern that they're going after? And then let's go back to the three underpinnings of the thesis, so to say the three parts, the scale of compute, the resilience of compute, and the flow of compute. If we just try and unpack each of those and say, if we start with scale, what are the technologies and solutions that fit within that part of the thesis?

29:12So the reason we're not saying we're a semiconductor fund, we're not a quantum computing fund or anything like that is because the technologies are somewhat temporal, right? They're accurate now, so I can name some technologies that are fitting today that we're looking at, but there's constant evolution here. So we can't tie ourselves down to one specific technology. Compute is a very wide concept that in some cases we can apply to very specific types of technologies, say semiconductors or photonics or quantum computing. But to name a very out there example, I'm talking to a founder that is growing brains in a vat and using them for compute.

29:59He can literally program brains, brain tissue, and use it for compute. The brain is sort of the most advanced compute device that we have access to. We just need to use it, right? So, okay. And then you would answer the same when I ask you with resilience. Maybe framing it slightly differently. What are kind of interesting developments or inflection points that we're seeing right now in these three areas? so for scale we still today have access to high-end fabs so in taiwan or korea or the us but as europe we're still pretty far behind on developing our own we actually have the knowledge i think we could build it but it would require massive massive investment what can we do instead is we can keep the or build the ip for inventing new types of compute one of course being this This brain example that I just gave, another one would be new types of photonics.