How does Fly Ventures approach frontier investments and support founders? Find out with Gabriel Matuschka.

Robin Haak: Hello, Gabriel. Thanks for joining us today. So, the first question, straight out of the gate. I read on your LinkedIn that you were thinking about the state of AI at this moment and whether we're at the stage with AI that the internet was in, in 1998. If we are, what are the reasons for that?

Gabriel Matuschka: There’s a lot to unpack there, right? This idea actually came out of a longer discussion we had at Fly in 2023 and then again during our AGM a few weeks ago. The question we were trying to answer for the group was this: if you take the analogy of the early days of the internet, there was a moment where things crashed badly before they became truly great.

Sometimes, we think there are patterns happening in AI, especially in marketing, that feel a bit similar to the pre-crash internet days. That’s things like $100 million inception rounds or companies being valued at multiple billions with very few - or even no - customers. I don’t want to overdo the comparison, and I’m not saying this will happen, but we felt it was our duty to think about whether we’re facing another crash-reset moment and what that might mean.

We also have a responsibility to our LPs to make sure we’re not on the wrong side of things if it does happen in terms of getting too hyped up. I also want to stress that it’s very amateurish to think you can time markets. We’re not trying to do that and won’t. We’re just using this as a way to think through certain things in AI that, in our view, don’t make sense for most funds to invest in, including us.

How does this then inform your strategy for investing in AI-first startups?

The way we look at it is that we try to either back the very first bet in a space or something logically close to the “last bet” and avoid everything in between. Let me explain what I mean by that.

The “first bet” could be something like Orbital Materials, which we backed a while ago. It’s an AI foundation model for material science, and at the time, almost nobody was interested in looking at material science meets AI. Sure, people realized it was an outstanding team, but it was still very ‘frontier’. The “last bet” means betting on things that have been tried several times before someone comes along with a very novel approach,  or "last bet" attempt. That was the case with Wayve in 2017. Wayve is working on an end-to-end AI approach to AV. Billions had been invested in AVs before and next to nothing had worked, but Wayve had a new way of doing things and has gone on to secure a lot of investment.

What’s nice about being early in a space like that is, first, you don’t see those crazy inflated initial rounds. And second, you don’t have 85 companies starting at the same time all doing the same thing just because it’s suddenly ‘consensus’ - that you can do this now. But that’s not to say we’re trying to do non-consensus things. It’s more about religiously hunting for things that other people aren’t looking at yet.

Let’s talk about that a bit. At Fly, you try to back founders who are solving hard problems and are focused on deep tech. What is it about this space that excites you, and what made you want to start a VC fund focused solely on this area?

So, the concept of being a first check for technical founders solving hard problems has a very, very strong “why” for us. The reason is, that I don’t want to tell my daughter—who’s seven today—that I spent 20 or 30 years of my life essentially backing companies that create automated spam messages to sell more Netflix subscriptions.

I mean, sure, there are interesting companies you can build in those areas, but we genuinely believe that if we don’t dare to tackle hard problems, we won’t progress as a society. We’re not an impact fund; we’re doing this to drive returns. But I think there’s been a bit of an overemphasis on replicating what worked well in other spaces—like scooters and food delivery—and applying that approach to AI. You see this behavior where money is pumped in to scale something really big, really fast. It’s all about raising rounds—A, B, C, D—as quickly as possible and getting on TechCrunch with flashy announcements. But, honestly, that doesn’t often produce companies where, if I look back, I’d feel proud of being the first to back them.

Take something like Wayve, a company that is pioneering a new approach to autonomous driving using end-to-end deep learning and computer vision. When we made the initial investment in 2017, we knew it

might not work, but if it does, it’s not just a great financial outcome—it’s something that could genuinely change the world. That’s the kind of work I really value. For us, there’s a strong “why” in focusing on this type of challenge because we feel the world has overdone it with what I’d call irrelevant problems.

And sorry for being so blunt here, but I think it’s a highly irrelevant problem to deliver Nutella to someone on the third floor as a matter of convenience. That’s not the future I want to contribute to.

What is the future you want to contribute to?

I believe—and I think there’s good reason to believe this—that societies thrive on productivity gains. One of the reasons Europe is so behind today is that we’ve failed to achieve those gains. We missed out on creating a whole series of very large tech companies here, which, in the end, is how societal progress happens. That’s what technology can bring us, and that’s what we collectively at Fly believe we should be focusing on.

We’re after those very large outcomes—technically difficult things, but if you get there, you’ve created something transformative. Something like OpenAI, not something like Gorillas with minor twists.

So, how did you get to the point where you’re investing in companies that will hopefully shape the future?

It’s been a lot of twists and turns—and, like so many, I got into VC a bit by accident. This was not my dream job. I pitched to many German VCs and I’m sure, I was a pretty horrible founder, I didn’t see VC as something I wanted to do. To put it politely, I didn’t think the customer experience was great.

Then, I ended up getting into VC and realized it was something I could actually enjoy—and maybe even felt a duty to do. You can complain about how Sh**** things are in Europe, or you can try to change them. For me, it became about contributing, even in a tiny way, to how this “product” is offered.

Specifically, for Frederik and me, when we started this, it was clear there was a niche in Europe: very technical founders raising their first rounds. These founders often only have their vision and past work to speak about—there are no metrics to show yet.

What we heard from these founders was that they felt they had to dumb down what they were doing when pitching, especially to people who didn’t have technical backgrounds. That’s not ideal for backing truly hard, technical work. So, with my disillusionment in VC and Fred’s technical background—we combined forces and decided, okay, let’s try to do something different and focus on inception-stage founders.

How did you meet your co-founder? How did you come together?

Frederik and I were connected by a long-term mutual friend, Stefan. Stefan and I studied together, and he was the lead PM for directions in Google Maps—so if you didn’t arrive at your destination for a while, that was his fault! Frederik and Stefan worked together at Google in Zurich. That’s how Frederik and I connected—through this shared, 20-year-plus connection with Stefan.

What do you both bring to the team? How do you complement each other?

We’re radically different people but with radically aligned values.

That’s a nice way to put it. So, let’s talk about raising Fund Three. What’s the process been like this year?

In general, I think anyone who says they raised in three weeks—with very few exceptions—isn’t telling the truth. It’s work for most of us, and it was work for us too.

We were fortunate to have made some bets in Fund I, like Wayve, that have done very well. If you get in really early and they raise something like a $1 billion Series C, it has a significant impact on a €35 million fund. With 9fin and another company, there are two to three companies of that caliber in our first fund, and potentially several in our second fund. That track record made our existing investors comfortable enough to continue the journey with us.

What challenges have you faced in bringing new investors on board?

We have a fairly concentrated LP base—there are only 12 investors in Fund II. So, raising Fund III was largely about insiders coming back. But it still required deep conversations.

It’s a tricky time, and for good reason, many investors are asking, “Is this a black hole where you keep wiring money and get nothing back?” That’s an issue for almost all of us in Venture Capital at the moment and something that we had to talk about with our investors quite a bit. We had to spend time explaining our approach and how we are addressing these challenges.

And deep tech often takes longer to see returns—it’s a longer game. How do you convince your LPs that investing in this space is the right strategy?

Honestly, I’m not so sure we do ourselves any favors by saying deep tech takes longer. I’m not convinced that’s true. Non-deep-tech companies also seem to take longer these days—they just face different kinds of challenges.

This ties back to what I said earlier. I think we’ve been polluted by the idea that rapid scaling - like what we saw with Gorillas or Bird is considered to be ‘success’. But maybe it’s actually a bad thing when companies raise too much, too quickly. For example, in what we do, it takes on average 21 months to raise the next round, which is clearly not rapid fundraising. But, maybe that’s a good thing. Maybe it’s better to raise at major inflection points.

So, to address the question that you initially asked, as for returns, small funds like ours can do partial secondaries in superbly performing companies before they IPO. In a €35 million fund, a $10 million secondary can move the needle for us without negatively impacting the company. At the same time, we’re honest with our investors—we tell them this is a long journey. Ideally, they’re ready for that and enjoy the process because they see what’s being created.

And what’s your average ticket size for investments? And, as a first-check investor, how do you source and find startups to invest in?

So today, from Fund III, we invest anywhere from €1 million to €4 million. That’s significant since we only do inception-stage rounds. This means we’ll invest in rounds starting at about €2 million, which is often just us and maybe one or two angels, up to rounds of €8–10 million, which do occasionally exist.

The vast majority of what we invest in comes from founder referrals. What we aim for is that someone might mention to a friend of ours that their colleague at DeepMind—who’s fantastic but a bit disgruntled—might be planning to give notice in six months. That’s where our involvement starts.

It’s often a longer process. You might discuss with them what excites them and why they’d leave a job paying $3 million a year for something with a much lower expected value. These are people who aren’t bad at math, so the motivation needs to be deeply intrinsic. We like getting to know founders as early as possible on that journey. Especially for more frontier opportunities, it’s often a six-month or longer process from first contact to wiring money. Some people think the inception stage means there’s no process, but there can be one. That said, we’re equally happy to move quickly—sometimes in seven days—especially for dev tools or infrastructure startups. Those companies often bootstrap, build something, or open-source it before seeking pre-seed funding. So, we’re comfortable with both approaches.

And talking about referrals and community, are you still running the Berlin tech meetups?

I’m not, simply because I don’t have the capacity anymore. I started those back in 2012 when doing something technical and demoing a product in Berlin was still a novelty. It was during the Rocket Internet heyday when all people were thinking about was which US peer model to clone next.

I created the meetups because I thought it would be a great way of advocating and showing off cool tech. But by the time we stopped, about five or six years ago, I didn’t think it was so necessary anymore. Also, running a pre-seed firm and having led 59 investments makes time a bit scarce.

And talking about those 59 investments, are you quite hands-on in terms of what you do with the startups?

Yeah, we have to be. Often, it’s just us as investors, or we’re the lead. But we’ve also learned that founders have very different preferences for how they want to interact with us. Some want rapid-fire interactions very frequently - say, on WhatsApp several times a week - while others prefer fewer but longer sessions. It’s not a one-size-fits-all approach - you need to cater to founders in a more boutique way. You have to adapt to what they need.

That said, there are key moments where a lead investor has a duty to act, like helping secure follow-on funding. And in our style of investing, while about 60% of our follow-on capital is from the US - where there are slightly more specialized VCs we will oftentimes have to help others navigate the funding maze a bit more. On the other hand, very technical founders don’t need us to help them design roadmaps early on.

That’s different from maybe 10 years ago, right? The type of support people need now compared to when you were running meetups?

It has changed, for sure, but it also depends on the type of bet. Some technical founders mainly want to spar on go-to-market aspects, while others with delayed go-to-market strategies have different challenges. It’s more specialized now, but the core dynamic hasn’t changed much.

Gabriel, what would you like to do more of?

I don’t want to write larger checks or invest in more companies - I think we’re doing exactly what we want. But I’d like to focus more on themes, like Germany’s industrial upgrade.

There’s this concept of “American dynamism” happening in the U.S., and I don’t think Germany, in particular, can afford to let more technology waves pass us by. So, I’d be willing to do more in this area. We’ve already made some bets along those lines, but as a German citizen, I’d like to see more of it. That said, we generally take a very pan-European view in what we do. And since we’re a tiny pre-seed fund, I know we’re not going to change the destiny of Germany. But I’d still like to help make it clear that people are willing to take bets against the current trends.

It’s not that I have the perfect solution for what needs to be done—because if I did, I’d just do it myself. But I’m willing to take certain risks. For example, there’s a company we backed in Munich called Software Defined Automation. They’re bringing CI/CD and other modern development methods, like backup and security, to programmable logic controllers (PLCs). These are essentially small industrial computers that manage machines and robots in factories. Germany has plenty of factories—dozens or hundreds of them with PLCs controlling machines. This kind of innovation is crucial for an industrial upgrade.

And when you talk about an industrial upgrade, what exactly needs to happen?

The truth is, a lot of the tech developments we see in other areas of our lives haven’t fully reached these industries. For example, today, you still have people running around with USB sticks to flash machines. It feels like the CD-ROM days of updating computers in the late ’90s.

And I don’t know anyone who’d say, “I’d love to go back to that—it’s much more fun than getting updates over the internet.” Yet, that’s still how things are done in many factories. This matters because industries change. Machinery and equipment are updated much more frequently now. So, beyond creating new machines, there’s a massive opportunity to manage what’s already out there more efficiently. Another example is a company in Munich called Robco. While we’re not investors, they’re doing interesting work in robotics. We are investors in another robotics company working on teaching robots to adapt to changing production lines.

It’s tricky to make the most of robots and help them do the work that humans ideally shouldn't be doing in the first place. But this kind of work excites me because it plays to Germany’s existing strengths, stuff that Germany in particular is good at. I don’t think it’s a smart strategy to say, “Germany needs its own OpenAI.” People have tried, with efforts like AlphaFold, but that’s not what I’m craving or focused on.

Now, let’s shift gears to some personal questions. What’s your background, and was there a role model who inspired you to focus on the future and the world’s challenges?

No, there’s no role model, and I’m not going to invent one.

I’m truly a tech enthusiast, and I believe technology can solve fundamental problems. I got into this early on, and a key experience for me was being part of the first cohorts of the CDTM program at TU Munich and LMU. That program, with just 22 students per semester, had a huge impact on me. It showed me what’s possible with a small, focused group of people.

That’s a powerful foundation. You also mentioned earlier that you had a glimpse into the VC world early on. Can you share more about that?

I had the chance to intern for Pitch Johnson, one of the first people to do venture capital in the Valley. He started in 1968—yes, 1968! There aren’t many people who started VC back then. I worked with him in Palo Alto for a summer, and it gave me this sense of what’s possible in venture capital. Pitch was part of the team that backed Tandem, which became Compaq and later merged with HP. They were also early investors in Amgen, which had a massive IPO in the pharmaceutical space.

Hearing those stories—seeing how early-stage investments could have such a huge impact—sparked my excitement for this field.

And you’ve bridged both worlds—being a founder before becoming an investor. What do you bring from your founder days into your work as a VC?

If I could go back in time, I’d probably tell myself I was a horrible founder. And I really mean that—you can write it, and I hope you do! But I think being a less-than-great founder gives me something important: empathy.  It’s funny, but sometimes just not being a jerk—having empathy—already sets you apart in this industry.

It’s easy to lose perspective when you’re in this position. You only raise money every three or four years, so you forget how hard it is. Then, when you meet founders, you’re the person writing the check—it can make people a bit detached. Maybe because I wasn’t a great founder myself and didn’t have the best VC experience, it keeps me grounded.

That’s a great perspective. Last question—what gives you balance in your life between your career and home life?

For me, it’s my family. I have a fantastic seven-year-old daughter and an outstanding wife, and they’re the best things in my life. One of the best ways to disconnect from tech and foundation models is to build a Lego train. It’s very grounding. I know it’s not the most magnificent answer you’ve ever heard, but switching between work and family—even in a two-minute time frame—is the most refreshing thing for me.

That’s a great way to balance things. Speaking of balance, there’s a lot of discussion now about AI-generated content feeling overly polished or manufactured. What’s your take on that?

We spend quite a bit of time on this internally—spotting when something was written by AI. Sometimes it’s obvious because it comes across as overly polished, and it lacks that human touch. If you’re writing for The Economist, maybe that’s your natural voice. But for most people, it’s clear when it’s not real.

Okay, last question! What would you say are your top three achievements?

I think what we're doing great is staying focused. A lot of VCs start great but lose their way because they get distracted by too many other things. So I'm most proud of our focus. Second, building long-term trust with our investors. That trust comes from doing what we say we’ll do—and following through. And third, and maybe this is my number one priority, being a great dad while being a great VC.

Thank you so much!

Thank you! I enjoyed it—it’s been a pleasure.