Lonis Hamaili: Launching an AI Supply Chain Hedge Fund

Brandon Beylo (0:00)

This is Brandon from the MacroOps Value Hive Podcast. Just a quick note that nothing you hear is investment advice, and you should always consult an investment professional before making any investment decisions. Everything that I say or any other guest that says anything in this podcast is for educational and entertainment purposes only. Please do not take this as investment advice. It is not to consult your professional. And with that out of the way, let's get into the show. No better time to talk about semiconductors and AI than today, right, Lonis? We are down. We're down. I mean, the DRAM ETF, memory ETF is down almost 30% from its highs just like two weeks ago.

And South Korea is imploding.

Cosby is down 7% a day last two days.

Lonis Hamaili (0:47)

Yeah.

Brandon Beylo (0:47)

But here you are launching a brand new fund, Cambrian Management, and the tagline is investing across the AI infrastructure supply chain. And so we're going to talk all about that, and where you're finding the best ideas, how you're structuring the fund, the principles around it, the thesis. But before we do all of that, let's get a background on you, how you got into the AI side of things and why you decided now is the time to start a fund.

Lonis Hamaili (1:18)

Sure. Yeah, great to be here. So I'm from Sweden originally. In 2019, I spent a few weeks in college, and then I quickly dropped out and started an AI company in e-commerce. This was pre-GPT-3, so it was more traditional AI, applying traditional AI in e-commerce.

The turbo markets, complete red ocean, low margins, and not good business, but we grew with a really good team. So we were able to grow within e-commerce and get a really good footholder, and go to more than 25 employees. And it's still growing, but we brought in the next almost CEO for some time to run it day to day. And then my co-founder, Anton, he went on to fund Lovable, and then our growth guy Max, he went on to fund Legora, which is this AI legal thing, worth 10 billion or so. And then our tech guy Oliver, he went on to fund Tandem Health, which also is a unicorn in Europe. And so we had a really good team. I think we were quite unique with being very early to recruit from programming Olympiads and physics Olympiads, and nobody was really doing that in 2019 in Europe at least.

And so I did that at some point, went on to fund the work, which was a crypto platform, and sold that in 2022, and then in 2023 started getting really obsessed with agents, you know, I'd be following AIS in 2017, but in 2023, it started getting good enough to actually perform some actions. And so we built a cloud co-work equivalent, but three and a half years too early, I guess in 2023

That helped people like use AIS to take actions on the web, and called Godmode, it got more than a million users in a few weeks. Most as a gimmicky product, right? It was a little bit too early still for technology, but I spent many thousands of hours on harness engineering and prompting, and trying to make models less dumb, which has actually become more useful now in the fund, where we spend a lot of our time building AI tools to help make our trading more better informed.

And then, I guess, a year ago, I also started thinking a lot about how to best position yourself relative to fast AI timelines, right? So if you think that AI is coming way faster than people think, and what the market thinks, what is the best way to play that? And it's not obvious, right? So you can take a job at the lab, we must try to invest in the lab, this is the end, and I did some of that. But it's not great because the government will very likely get involved in many of these labs. Right? And so it's not clear to me that the labs can grow to more than 10 trillion, let's say, before government gets involved and really kind of hurts the enterprise value.

And also there's Chinese risk and blah, blah. You can also take a job at the new clubs. But then again, same thing, execution risk, who wins, etc. And so I came to the conclusion that actually the best way to play this thesis is to be very long, all the important layers of the supply chain. Because whoever wins, they will need the same memory, they will need the same power, they will need the same ships and underlying tools. And so that's basically the thesis. And I guess I started catching up with my friend, Kuram, who had actually been trading on this thesis for four years or other. So before any of the other funds you might have heard about, he was trading on this. And he had returned something like 4,000X on his own portfolio. So he had like had quite outsized returns just trading on this. And so I joined him, I guess, started trading personally and we both did really well. And then I guess three months ago or so we start, we start talking to friends and all of them were basically looking for someone, AIPL, the technical to manage their money with this thesis. And so that's how we started basically, I guess we've been seeded with around $70 million and started trading I guess in September.