Quantum Light invests equally across 30 to 40 startups

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

1. 49:26

   It's basically. You also set up your own, I don't know if you call it a VC fund. Yes, it's a VC fund. What's it called, first of all? Quantum Light. Explain what you're doing. This is funny. I just picked up on this now because you were analyzing the legacy banks. You're like, I'm not. Look at all the obvious mistakes they're doing. Some of that came to experience, very hard-fought experience. The way they're doing it doesn't make sense.

2. 49:51

   I'm going to show you a different way to do it. Yeah, basically the idea was in, I think it was 2022 or 2023 when I came up with this idea. And the idea came actually from my personal pain, from my fight with the investor. It was actually one of their big names. They entered the investing revolution. It was, I think, just after COVID or during COVID, when all the stocks went up. They invested at $33 billion valuation.

3. 50:21

   So basically, they had legal right not to allow us changes in shareholder agreement because they were like, you know, as an investor, blah, blah. So you had a bunch of these bad experiences with VCs. You're going to start your own VC fund. Explain how your fund is different. So basically I had this fight and then, you know, I was extremely pissed off and I thought, you know what, I also need to compete with VCs in their business because they're so bad.

4. 50:45

   And out of my frustration, I started the fund as well. And then the idea was basically to invest not based on the human judgment, but based on their science and based on their machine learning models. Statistical investment is number one. Number two, actually help with founders with operational advice, playbooks that we developed at Revolute, how to hire people, how to do performance management, how to do product expansions.

5. 51:09

   And then finally, help them sell their product as well, because we are a large enterprise. We are buying from a lot of vendors and having us as a client helps to sell into other enterprises or potentially sell to our customers as well, our Revolute customers. So I think overall, plus my personal advice, obviously. So I thought overall the package was so much better compared to purely financial players, VCs.

6. 51:37

   You just said your personal advice as well. So like, are you just making concentrated bets? Is this like another thing where you're doing 20 or 50 at a time? Or are you just like concentrating money behind a handful? It's again, it's statistical investment, right? So it's based on machine learning models. So the way it works, our model selects 30, 40 companies to invest over two, three years. and then puts equal amount of money in each startup.

7. 52:00

   So we started with a fund, a Series B fund. So it's closed, it's fully invested. It produced now like 56% IRA. And you started like $250 million or something, right? We started $250 million, yes, with the first fund. So now it's like 56% IRA. We started investing in second fund now, which is $500 million. So it's Series C now in companies. Again, based on the model, also very high return. I think it's top 5% of fans in the world.

8. 52:30

   now in terms of returns we'll probably do a large fund as well for for leads where's the money coming from it's different lps so initially i started 20 was my own money so i put 20 of it or 25 of it and then the rest came from other founders second fund we already got quite a few institutional place as well i've never heard somebody building a bank compare themselves to like meta so i'm gonna get into that in one second but you don't do anything like half-assed or like small right It's just like, okay, I started at 250.

9. 53:00

   Now you're just gonna see how far you can take this. Is it gonna be like a giant, like renaissance technologies kind of thing? Well, obviously the goal is to build something large like BlackRock, but in private markets. Again, we'll see if we get there. But I think idea and approach makes sense because like you now have a lot of data on startups and then you can apply all these machine learning models to identify the best ones.

10. 53:30

    And then model is pretty sophisticated. So, for example, one of their signals we'll look at, we'll look at who invested in the company, who is investing. And we'll look not only in the fund performance who invest in the company, but down to the partner level. And we'll also look at the age of the partner. And there are very interesting things when you look at these variables as age of the partner.

11. 53:52

    And then you'd rather not invest in companies where partners who are investing are older than 40 years old. So the best cohorts are 30 to 40 years old of young partners who are investing. They're the best at identifying good companies. Same as with executives as well. So there's a certain prime age, what you need to look at. Again, there are many variables, right? Age is not only one thing, right?

12. 54:18

    But statistically speaking, that's what you need to look at. And is this a separate company from Revolut? Yes, completely separate. Okay, so completely separate. And you have somebody else running it? Yeah.