Sasha Vidiborskiy, Atomico: How to Think About Deep Tech Investing

0:00You don't only need to prove that quantum can be faster, you need to prove that it can be faster for tasks that are important at scale. What we mean when we say deep tech, whether it's software or hardware or a mix of both, it's the sheer belief that the research that the team is doing can compound, can allow them to create a very unique solution that is 10 times or 100 times better than the next misalternative, and it's not easy to quickly replicate. So one of the things that we're testing for is, yes, you are extremely technical, but where are you coming from? Are you coming from, I found this problem very cool, or I always dreamed of solving X?

0:35Or are you coming from the point of, I want to build the biggest player in this space because I know that in order to make the biggest impact with my technology, I need to be everywhere. And if I'm not commercially successful, no one cares. Eventually, the entire deep tech story is about simplifying complexity, to my first point about why complexity I believe is so important, and telling the story in the right way to different stakeholders, employees, getting them excited to join something that is going to fail with 97 % probability.

1:02Andreas Munk Holm:Welcome to the podcast, everyone. Today, I am super excited because we're sitting down with Sasha Wieryboski, who is a partner at Atomico, a London-based venture capital firm with over 5 billion assets under management. And they, of course, invest from seed to growth across Europe and beyond. At Atomico, Sasha focuses on deep tech and has led investment in companies like Lace Litography and La Cara. Before joining Atomico, Sasha was a senior associate at Runa Capital and earlier started his career in quantum physics research, which I think is the perfect segue into today's conversation because today it's all about how Sasha thinks about deep tech investing.

1:40This show is not investment advice and the hosts of this episode may be invested in the funds and companies featured.

1:45Andreas Munk Holm:I said we're going to talk about all about deep tech investing. And to those that know me, I think we'll also know that I'm going to give you a bit of pressure here because I'm always saying, hmm, deep tech investing, generalist, there's a lot happening. Can you really be a generalist investing in deep tech? You, of course, are no generalist, but Atomico is. So that's what we're going to explore today, how someone like you are shaping how Atomico is thinking about a new and growing space in Europe. First and foremost, talk to me a little bit about the importance of your journey from being a quantum researcher to now being an investor.

2:25Yeah, I think it all started even earlier than that. I'm originally from Moldova, quite a small country and squeezed in between Ukraine and Romania. And I think as one of the former Soviet republics, there were two paths you can choose. You can either go and do what is most of the people do there. You know, it's agriculture, it's services, it's banking. Or you can try to go and do hard science as a way, hopefully, to manufacture for yourself a social lift. And for me, I definitely gravitated more towards hard sciences than farming. So I decided that I would do something that sounded cool at that time.

3:08And by first background, I was actually a material scientist with a focus on materials for electronics and then got pulled into what eventually became quantum computers. But unfortunately, 15 years ago, quantum computing wasn't a field that many people could get excited about. I don't think that anyone called it quantum computing. So obviously there is a bit of storytelling here in terms of what the industry eventually became as the components of quantum computing have matured. And now, you know, you even see like companies going public and being public in quantum computing.

3:44Andreas Munk Holm:I literally two hours ago did a big conversation with Andy Lieber from Notion on quantum. And you're nodding here because you know him, of course. Yeah, and it's like the fun fact is we still don't have yet a fully functioning and kind of utility scale quantum computer. Anyways, I think... Let me just get your take on this. Andy believes that, and he quoted Google for this, but he believes that we will probably have quantum at scale in 2029. And then we will have a conversion of quantum and AI being applied together in the early 30s. which is the time at which he believes we're going to see a massive unlock.

4:29Andreas Munk Holm:And that's where if we don't get AI sentience at that time, because it's just not baked into the models, we will have something that feels like it because AI can move so fast because it's able to compute at a completely different rate that is completely unfathomable today. Yeah. Yeah, I think quantum computing timelines is this kind of questions where if you ask someone at any point in time, they'll give you the answer that is five years away. Whether it's five years from here or three years or 10 years, I personally feel it's very hard to say. Why? Because I think the goalposts of what quantum computing utility actually means is moving.

5:12We invested, and this is to your point about Atomica being a generalist, we have been investing in deep tech for over a decade, including some of the early bads in quantum computing. So we invested in a company called SciQuantum, which is now one of the biggest private quantum computing companies. And we invested in 2018. So that was like eight years ago. And if I look back at our original memo and just, you know, contextualize it, definitely we were right on many things, but we were wrong on the timeline. And that's probably a cross-cutting theme in deep tech, just appreciation how long things take.

5:45So if I go back to quantum and the timelines for it, the goalpost has been, can you show quantum supremacy? Which basically means, can you show that quantum can do computations faster or more efficient than another class of compute, typically CPU or GPU, for a specific class of tasks, right? And there will be like benchmarks and so on. And Google proved that quite a few years ago. It was like five years ago, four years ago. A core piece of our investment in PsyQuantum and the teams believe is you don't only need to prove that quantum can be faster. You need to prove that it can be faster for tasks that are important at scale that is important.

6:25And what they say is they are going after not just a quantum computer, they're going after utility-scale quantum computer. So if you ask me about when will we see quantum compute being useful, yeah, maybe it's the next three years. But when would we see a huge cluster that discovers new materials or helps with encryption or optimizes hedge fund portfolios? I think that's probably a few more years away. And I'm a bit hesitant to say how many years, but probably a few more.

7:01Andreas Munk Holm:And I think that, honestly, by diving into our conversation about quantum here, we made it clear to everyone how your journey has shaped your investment framework today. Anything you want to add to it, though? I think for me personally, as I was investing and as I was thinking about what resonates with me, but also what resonates with founders, because eventually as an investor, we are serving founders. and you need to be very mindful that there will be types of founders or types of opportunities for whom or which you can be a better fit. And therefore, you have higher likelihood of winning their hearts and minds, but also being more useful to them versus other founders and other opportunities.

7:48And for me personally, as I spent quite a few years in academia, grinding pretty cool, but challenging science, For me, that ability to navigate complexity has probably become a core theme of what I look in founders, what I personally get excited about, like sheer complexity and ability to build something that is a simple solution for very complex problems, but also ability to communicate that complexity. So to your question around how it shaped my journey definitely has shaped what I tend to spend time with and who I gravitate towards. The best founders that I've worked with, they work on the biggest problems.

8:31And those problems in many instances are multifaceted. This is not something that you can solve with one specific product, one specific approach. And that complexity is probably my thesis, both in terms of why I invest in certain companies, why I get excited about certain companies, how we navigate that complexity, how we can help founders to get that complexity, to get through that complexity, and how as founders, you need to be communicating that complexity so that the world understands. So to summarize it, probably complexity stemming from my experience in academia is the theme that I kind of in my mind keep at all times, pretty much, you know, every day that I do work at Automiton.

9:19Andreas Munk Holm:Satya, I want to try and tease out your framework for investing in deep tech. And the reason why I think this is relevant is because everyone, a lot of people are painting themselves as deep tech investors today. and deep tech investing can mean many different things. Some would say, well, I'm investing in an AI application thing. It's using machine learning. Machine learning is definitely cut off the, you know, and for that reason, I'm a deep tech investor and that's what I'm going to go out and say to the market. And others would say, I personally love the IQ capitals approach where they're saying if you can bring together five of the best people in the world in a space and they can build what your solution is doing in three years or five or whatever is the number they use, then it's not deep tech.

10:05Andreas Munk Holm:I like that. It's very, like, we can all get it. We can all understand it. What's your take? It's a great question. I remember vividly we had one IC for a company that is in the developer tooling space, like it's an infrastructure for developers. And I remember in our IC discussions how, So because it was a very complex product and it felt very deep in the developer tool in stack and the kind of infra stack, people by default were asking, does it mean that they have a moat? Because, you know, it sounds complex and therefore probably it's harder to build. I inherently don't believe that there are pretty much any moats in the tech itself.

10:52Anything in my mind can be rebuilt with proper resourcing, with proper motivation, which is a critical piece and proper amount of time. Even if you think about the most sophisticated things, they can be built with these three things in mind and at hand. Obviously, patents and other forms of protection help. And there are entire industries like pharma and so on that have operated on this. But in startup world, it's very hard to build something so unique that no one else will be able to build it. Obviously, now it gets to a question like, what then can you do to make your approach more defensible, which is a separate question that I'm very happy to dive into.

11:32But if we think about how we define deep tech, we define deep tech as an industry or a set of industries and verticals and approaches where investment in R &D can fundamentally make something more defensible rather than less defensible. And it's something that compounds over time. So in practice, what it means, if you look at the P &L and balance sheets of like SaaS companies versus deep tech companies, SaaS companies also spend a lot on R &D, and we're talking about anywhere between 15 % and 30 % of their OPEX spent on R &D. If we look at deep tech, we're talking about 30 % to 50%. And then the most important piece is not only spending on R &D, but seeing how that R &D spent compounds and allows you to leverage that R &D spent to do something.

12:20It might be distribution. It might be product building. It might be network effects. It might be something else. But eventually, you want to see that what you invest into allows you to build deeper defensibility. I don't, once again, believe that it's a mold in itself, but it makes something much harder to develop. And if you think about some of the biggest examples like Apple, no one thinks about Apple as a deep tech company, but it's a deep tech company because why cannot a random company with a lot of capital replicate what they have, because they have invested a lot into the processes, into the R &D, into the patent portfolio, and a bunch of other things that allow them to stay ahead of others.

13:00On the other side, it might be like Saudi Aramca, right? Once again, oil and gas. You wouldn't think about oil and gas as R &D industry, or R &D kind of focused industry. But once again, for them, the ability to extract oil and refine it and transport and sell is based on multiple years of hundreds of billions of dollars invested. So for us, what we mean when we say deep tech, whether it's software or hardware or a mix of both or something else, it's the sheer belief that the research that the team is doing can compound, can allow them to create a very unique solution that is 10 times or 100 times better than the next best alternative.

13:42And it's not easy to quickly replicate. It can be replicated, but not probably as quickly.

13:48Andreas Munk Holm:Let me just double click on one thing because you said 30 % to 50 % of the money is spent on R &D. This is at the seed stage? Just to clarify to anyone that's listening. Yeah, it will be there. I appreciate when I talk about numbers, I need to be specific what they refer to. I think 30 to 50 percent, if you think about scaling company, and we're talking about potential hundreds of millions of invested, maybe that is a good mark. That's where you start investing a significant portion of that capital. Obviously, as a proportion of your spend, it might be bigger or smaller in the early days, depending on what the company is focused on.

14:30If you're at seed stage or five people and you have like three million budget, so to say, of course, your scientific R &D can be a big proportion of that can be like 70 percent, 80 percent. If you need to buy one tool, right, that would be R &D spend, CapEx. So from that perspective, it's kind of the devil is in details, but it's more directional. How are you thinking? Are you thinking about, you know, go to market?

14:54Andreas Munk Holm:It's more when you're thinking about how the company will evolve over time. Exactly. What trajectory are we on? Are we on one where it's distribution that's going to be giving us our moat? Or is it one where it's the research and development that's going to build it? That's exactly right. Okay. I wanted to clear it up in case anyone was thinking, well, blah, blah, blah. Like, yeah, don't try and be devil's advocate against someone who's trying to explain a complex framework here on a podcast. I want to ask you then, second thing here. it's a question that's close to my heart and it's what founder attributes do you believe are really important when you're investing in deep tech companies and also how do you assess assess that and I asked that question with great curiosity because I used to actually be a bit of a chief of staff type role in a deep tech fund and we kept the fund was incepted by the government in Denmark to basically help those companies that were spinning out of universities, that were not getting funded by the private market, and then there needed to be this actor in between.

15:59Andreas Munk Holm:Very often, it was founding teams where there was lacking a commercial profile, and then a commercial profile would be brought in. And I kind of always cringed a bit at that because I was like, you know that that has to sit with a founder. You also know, and I think we've learned that a lot in the past years, here in Europe, the importance of having the CEO be very technical, also in terms of being able to talk to customers. We just see a very close connection between the technical expertise of the CEO and the success of the company, while still the CEO obviously needs to be commercial, but oftentimes it's very difficult to take someone without technical knowledge, deep technical knowledge, and put them in as CEO of a startup?

16:49So the way we think about founders in the context of deep tech, obviously there are qualities that we look for in any founders, right? If you think about resilience and drive and passion and so on, that is almost like a base layer. I think deep tech, and that's where the nuances are important, deep tech brings additional flavor to the company journeys because in certain moments of company formation and pivotal moments for the companies, there will be things that will be critical. And if founders don't have certain skills or traits, it will be harder for them to go for those pivotal moments. Maybe just to start with one that you touched, which is the commerciality.

17:33I think the reality is if you are in a very complex and technical industry, it's more rare than not that you would have founder skill on the technical side. And that is fine. To your point, you need to be very technical. If you're CEO of a company that pushes the frontier, and we're not talking about marginal improvements, we're talking about leaps in terms of whatever you're talking about, performance, it might be market sizes and so on. You need to be technical, but you also need to be commercial. And what I mean by commercial, you need to have desire to build a company, not to necessarily build a cool product or solve a technical challenge, but to build a company.

18:21And building a company, if you're a for-profit company, means building a commercial engine that will eventually make that company successful. So one of the things that we're testing for is, yes, you are extremely technical, but where are you coming from? Are you coming from, I found this problem very cool, or I always dreamed of solving X? Or are you coming from the point of, I want to build the biggest player in this space because I know that in order to make the biggest impact with my technology, I need to be everywhere. And if I'm not commercially successful, no one cares. Once again, many caveats and so on and so forth, but you need to have that drive and desire to build a company.

19:04And that pretty much clearly comes through in the first couple of interactions that you have with founders, because when you ask about their outcomes that they dream about, if you ask about what they're prioritizing different times, you would hear, are they prioritizing what will eventually make a successful company? Or are they prioritizing something that they personally feel they need to do? Solve a problem, build a product, and so on and so forth. So that's number one. The number two is in deep tech, it's probably the most unexplored territory ever, because for most ambitious companies, you're doing something that either no one has done before or people think it's dumb to be doing.

19:44And you kind of need to persevere and push through and explore what can be possible and push that boundary. And while you're doing that, there are so many things that you can be doing that you need to actually know what to focus on in different times of company building. Are you now focused on like one customer? Should you expand to 10 customers? Are you focused on one application of technology or should you expand it to five applications? Are you focused on building the product or talking to partners? And so on and so forth. And I know it sounds somewhat similar to what a typical company would experience.

20:19but with deep tech, because in most of the cases, you are building the frontier. And building the frontier means that no one knows what it is. So it's very hard to know what the signal is and what is the noise. So having that ability to focus and not be distracted by the things that will eventually not be important is paramount. And it's not easy to test for because in early stages, by default in constrained environment, people are focused one way or another. So the question becomes, when they get that capital, when they get more time, are they going to be as focused? The third one is probably a piece that I already mentioned, which is desire to build something that is 100 times better than the status quo, rather than, oh, we're 30 % better than the next best alternative.

21:06Don't get me wrong. If you get 30 % more efficient motor, it still might be great. but not necessarily the type of investments that we're looking for, and also not necessarily the type of ambition that founders would have in pushing that frontier. And the last piece, which I think is very rare in technical founders, is ability to storytell. You mentioned you need to be speaking to your customers and to other stakeholders. Eventually, the entire deep tech story is about simplifying complexity, to my first point about why complexity I believe is so important, and telling the story in the right way to different stakeholders.

21:47Employees getting them excited to join something that is going to fail with 97 % probability. Partners who need to see the product, but you don't have the product and you want to convince them that you're the right company to spend time on. Investors, obviously, fundraising is absolutely important as it is everywhere. But in deep tech, you kind of run out of money sooner than you think. So therefore, being able to inspire people is critical. Yeah.

22:16Andreas Munk Holm:Now, I want to take you into understanding Atomico's deep tech institutional framework, so to say, in other words, how does Atomico deal with deep tech? But let me just preface this with something that might make it more accessible to founders to understand why is this interesting? Why should I understand how Atomico thinks about this? I know quite a few deep tech investors that are either inside big firms or have been inside big firms where it can be really difficult because if you're in a firm that has always done B2B SaaS or something like that or primarily done that and then you come as this deep tech partner then you're automatically in an IC and investment committee battling against people that are used to look at different metrics and are used to hear different stories, used to oftentimes also see more presentable founder types, if we should be so corey to say it like that.

23:14Andreas Munk Holm:And they love that they can say, no, no, look at it. My stats are this. It'll grow like this because that's what every other B2B SaaS company has ever done in history, and it's going to be like this. And that's kind of what you're up against when you're then saying, yeah, but I got these three cool guys that are building this in the garage, and it's going to be amazing because we're going to get free energy. And for that reason, it's super important as a founder that you understand the dynamic within the investment firm that you're talking to. Because sometimes you might have a partner that's great, but that doesn't mean that that partner has an IC that necessarily understands you.

23:49And they might love the idea at seed stage, but when you need to do a seed plus and bridge

23:56Andreas Munk Holm:to the next round, that's where it gets really, really difficult for a B2B SaaS investor because the existing knowledge in traditional venture has always been, if you don't make the progress that you need to make, you're out. But in deep tech, oftentimes that needs to be taken with a grain of salt. Comment on what I said, Sasha, if you agree, and then tell us a bit about how Atomico thinks about this. I'll probably start with the last piece because I think it's important. In my view, regardless of whether it's deep tech or consumer or B2B SaaS, you need to show progress. And therefore, the progress definition or the definition of progress might be different, but there has to be some progress to get people excited, especially if you're seeking new capital.

24:43Getting back to the overarching premise of generalist deep tech investors and then deep tech inside of generalist funds, I think every fund is different. And my biggest advice to founders would be make sure that the partner that you're working with is well equipped to navigate the intricacies of their own fund, because there are funds that are very understandable. There are funds that easy grasp things. There are funds that would be on the other side of the spectrum. Eventually, you want to start building a relationship with your partner in a way to help them build that case internally, if that is needed.

25:19I think it's not always needed, and I'll cover in a second how it works at Atomica. In my mind, because we've been investing for over a decade in deep tech, from quantum computers to flying vehicles to chips to lithography companies and so on, we institutionally have gained experience for a number of things that go right, that can go wrong, what to look for.