Mark Thomas announced the launch of his “AI-first hedge fund” Badass Capital in May, as the former ZenSports boss looks to bundle decades of industry experience with AI agents to gain edges across sports betting, crypto trading and prediction markets. ZenSports, which was licensed in US states Nevada and Tennessee and sold in 2022, was co-founded by Thomas with a peer-to-peer vision.
Now, with the support of AI, he is looking at the sector through a different lens. The rapid rise of prediction markets has also given him an injection of enthusiasm, while also hoping to harness the growing pace of change within the technology. Here, Thomas speaks to EGR about what makes up an AI-powered hedge fund.
EGR: What was the thinking behind the new AI fund and why is now the right time to launch?
Mark Thomas (MT): I’ve been on the other side of the table. We were a mobile sportsbook operator, so I’ve seen what that’s like. I’ve been betting on sports for 30+ years, I’ve been trading crypto off and on for the last nine years and full time for the last couple of years.
I looked at the landscape and tried to understand where the future was headed with regard to both sports betting and trading in general, and the proliferation of AI. The future of trading and sports betting has got to be with AI, and then obviously with the proliferation of prediction markets, I think really answers the ‘why now’ question.
I have started many technology startups, and I was always the non-technical co-founder. Now with Claude code, I can be the technical co-founder as well. I even built our website, which I had never done before, and of course, the AI models and agents that I’ve created. This is the future of trading and sports betting, in my opinion.
EGR: Why did you decide not to move back into B2C following the exit from ZenSports? Is it too tough a sector?
MT: It’s a great question. The marketing and the legal is a lot of work and money. If you look at all the big operators that are out there, they spend nine figures a month sometimes on marketing alone, and that’s hefty to say the least. When you’re a startup, that’s just not on the cards. When you’re a startup, you want to be able to do things lightweight, easily, and build and scale and grow.
My take in general is that I think with prediction markets, especially for edge betters, I think it’s just a better setup, a better model, because traditional sportsbooks don’t really want edge players. I think there’s a huge opportunity compared to just being a traditional sportsbook, where you have to spend a ton of money on marketing and a ton of money in legal with pretty razor-thin margins. That’s just not as attractive as I think what I built.
EGR: Can you explain how the AI agents and models will allow you to have an edge as a hedge fund?
MT: It’s two things. It’s time and it’s accuracy, with maybe a bit of more emphasis on the time aspect. Both trading and sports betting; it’s a lot of research that sucks up a lot of time. The second thing, of course – and without this it wouldn’t matter anyways – is the accuracy piece. I’m approaching 60% on the sports betting side for even money bets. I’ve basically taken all of my experience sports betting, trading, being on the operator side, and distil that into models that run automatically, even when I’m sleeping, and are basically constantly spitting out the right trades to take and the right bets to make.
I don’t yet have the agents or models actually placing the trades or bets for me, mostly because of risk management. Agents can go rogue, they can hallucinate, drift, etc. So, when there’s real money out there, I don’t want them doing that. My hope is, in 12 months, that part is really buttoned up, that they have context and they’ll execute the trades and the bets for me.
EGR: Does the eventual democratisation of AI reduce the sustainability of an AI-powered hedge fund?
MT: Will the edge eventually be lost? If everybody has a great AI agent, does it really matter at that point? A couple of things … It will take a while to get there, like seven to 10 years, so I think we’ve got at least a while before that starts to happen.
Secondly, some people are absolute AI power users, they do literally everything with it, while others use it sparingly. I would say anybody who’s sophisticated will have their model set up. I do think the AI agents are still only as good as the people who create them and what they input into them. If somebody’s just not that good, I don’t think their agents will be that great either. So, I think it’ll be like 20+ years, maybe 15, before you don’t have to do anything. It’ll be interesting to see what happens, but I think it’ll be quite a while before any ‘edge’ is lost, where anybody can just do it without really any kind of background or experience.
EGR: How much of a cannibalisation threat are prediction markets to online sports betting operators? And can the two sectors exist side by side?
MT: The recreational player who’s in a legalised state is definitely gravitating towards traditional sportsbooks still, whereas it’s obviously the states that don’t have legalised sports betting is where players are heading towards prediction markets.
I think it’s more than fair to say a traditional sports betting app nowadays still has a much better user experience for a sports bettor than a prediction market does. Can the prediction markets eventually get there? Sure.
I think traditional sportsbook operators still have a pretty big edge over prediction markets but I do think prediction markets can get there. It’ll be interesting to see how quickly the prediction markets catch up in that area to really invest in user experience and product.
While they offer a lot beyond just sports betting, within sports betting they don’t have a wide menu of sports bets to take, it’s just basically over under, money line and point spreads. That’s an issue, in my opinion. I think the prop bets are really light and, certainly from a liquidity standpoint, a lot of those prop bets have very light liquidity to them.
EGR: Do prediction markets end in shark-on-shark violence as alluded to by BetMGM CEO Adam Greenblatt?
MT: I don’t think so any more so than traditional trading is. You could say Robinhood would theoretically be that, but it’s not. People who are not sharp trade there all the time because there are depths of liquidity there. There’s still going to be recreational bettors who want to get real size down that may not be able to with traditional sportsbooks. There may be types of bets they want to place. There may be a situation where they want to save on fees or want to become the maker. I think people will become savvier to all of that as time goes on.
People understand what that’s like, and they understand there are benefits to betting or trading like that. So no, I think retail will use prediction markets as they see need for it – it’s not just going to be shark-on-shark.
I also don’t think recreational players are only going to use the sportsbooks. I do think from a purely user experience standpoint, in the states that have legalised sports betting, sports bettors generally gravitate towards the sportsbook operators, just because it’s cleaner and easier and faster for them to use.
EGR: Where would you expect Badass Capital to be in 24 months?
MT: That last mile that I mentioned, where the AI agents are actually placing the trades and placing the bets, is really where I want to be for all the sports, for all the leagues, for all the different trading types. Some of the models for the additional sports leagues, like the NFL, which hasn’t started yet this year, those still have to be developed. That’s going to be done by this summer. And then just continuing it, because each league and sport has to have its own agent. I expect in 24 months, every major sport will have the agents, and then I want them to be doing that last mile. We should just be sitting back and observing and just making sure nothing goes haywire.
The post Q&A: Badass Capital on AI-powered hedge funds and the liquidity issue with prediction markets first appeared on EGR Intel.
Mark Thomas, former CEO of ZenSports, explains to EGR the thinking behind the launch of the Miami-based fund and how AI has turned him into a technical co-founder
The post Q&A: Badass Capital on AI-powered hedge funds and the liquidity issue with prediction markets first appeared on EGR Intel.