“Markets Don’t Lie?” Kalshi’s CEO on Predicting the Future

Yeah, well, first of all, thanks so much for having me, Ari and others. So the story of Kalshi: we started at the end of 2018, but the idea was formulating for a while before then. So I went to MIT, and my whole thing was just, I wasn't a very social guy. I was a little bit of a math nerd, and by that I mean a super big math nerd. And I just always was enamored with the objectivity of math. There was something beautiful about probabilities and numbers, and it was something that I just felt more comfortable with. I went to MIT, and then it was like, okay, where are all the other antisocial math nerds going? And they were getting hired by finance companies, and I was like, okay, I should go there. And I went.

The initial thing was, I worked at Goldman in 2016, and there were two things happening that summer that were pretty consequential. First is Brexit — the referendum, I think, was happening in July 2016. And then second was the first time Trump ran for the election. And you know, you're sitting, typical Wall Street, at Goldman. There's hundreds of traders, and they're trading a bunch of weird stuff, you know, like swaps and calendar options and all these names that nobody really understands. But the basic thing was, you had all these fancy customers — hedge funds, mutual funds, high-net-worth individuals, family offices — and they were just asking, hey, what is going to happen with Brexit, and how should we play it? And then the second half was, what is going to happen with Trump, and how should we play that? What is the Trump play? What is the Trump trade? How do we hedge against it? And it really marked me at the time, because what we sold — and we sold this in droves — the big Trump trade that Goldman came up with, this super genius trade, was essentially short the S&P on the week of the election.

It was a bad trade. I mean, it was a truly, truly terrible trade. You could not think of a worse trade. The S&P rallied the most ever in its history that week. So all these guys were right about their prediction, they lost a lot of money. Goldman made money because we sold it to them, so we're not exposed. And I was like, we could build something better.

But the more I thought about it, the more I really got obsessed with the idea, because you have financial markets to price a lot of things — stocks, companies, how a company is going to perform. You have financial markets to price tangible commodities like metals and oil, and then intangible commodities, so that really started at the turn of the century — interest rate swaps and the stuff that really nobody understands. And really this was about, how about we build something that can price simple questions about the future, like will Brexit happen, and have people trade it, buy it and sell it the same way they buy a stock, and it settles — the way you get paid out is if it happens you get a dollar, and if not you get zero. And people are buying at any price in between, so if you're buying at 50 cents, that means you think there's a 50% chance of that event happening.

And I got into this theory around prediction markets, and there was some theory. So it really started with this economist Hayek in 1945, who talked a lot about this idea that information is surprisingly dispersed — it's surprisingly non-central. And we often make decisions in a centralized way. So you have a group of people, like a government or a company, that has a very small percentage of the information they need to make a decision, and they make some decision. It's often an okay decision, but if they could have all that information — that could be sitting in Hawaii, or in New York, or in Queens, or anywhere — they would make better decisions. And the best way to bring that information is via markets: incentivize people to come and trade and participate, to bring whatever information they have about something. Some people may be following Brexit and figuring out how their neighbors are feeling, and they may come and trade it. Some people are, I don't know, Trump fanatics, and they're thinking about how their neighbors might vote, and they might come and trade it.

And I really love the idea, because if you could build that market, one, it could be bigger than any of the other markets, because more people will care, right? And you're seeing how much people care about prediction markets now, because you get it. It's not very hard: will it rain tomorrow? How much is it going to snow in DC? Will Trump raise tariffs again? Who's going to win the Super Bowl next week? And the second thing is it'd be very useful, because if you believe in markets, they're a very good way to aggregate the crowd's wisdom to answer a question. And so if we could build this thing, it would make us a little bit smarter about the future, instead of like, 'climate change is real, 100%, everybody else is an idiot,' and 'climate change is a hoax, 100%, it's a hoax, everybody else is an idiot' — we'd come somewhere in the middle. Markets tend to not lie. They're not as polarized, they're not as aggressively subjective.

I didn't start a company then. I didn't really think about being an entrepreneur. I was really going to go work at Citadel and be a trader. But then the idea wouldn't leave me. I just realized after a year or two, if I don't start it, I'm just going to regret it for the rest of my life. And so we decided to start, and that's when the whole journey started.

When we first started the company, we were like, okay, why does this not exist? It seems too obvious. We can't possibly be the first that came up with the idea. And it was true — we weren't the first. A lot of people had come up with the idea before, but it was just not allowed in the US. And so we just decided, okay, if we're going to do this and leave everything and go for it, we are going to have to do it by the book, regulatory first, and we're going to have to legalize the space — which is a little bit of a dumb thing to think about at the time, but we were a bit naive, and it helps to be naive, because you just don't realize what it was going to take to actually make this happen. But we were so deeply committed, because one, we never thought we could get to proper scale. If it wasn't done within the right framework, people wouldn't trust it, it wouldn't go mainstream, and if it did, at some point regulators would shut it down. And two, it's a bit more philosophical. I always thought, what are the companies that I look up to? And they're companies like Schwab, and the new generation like Robinhood, Coinbase.

And it was pretty tough, because we started — we're like, okay, we're going to get it regulated first. And imagine two 22-year-olds calling lawyers, whoever would pick up the phone. In one day we called literally 65 lawyers, and all of them said no, until we found this one weird guy, like Jeff Banman. But he said, yeah, I mean, I was a regulator at what we thought would be the right regulator, which is called the CFTC. So the SEC regulates equities, this regulates everything else. And, let's just try it out. And imagine the first time walking into this financial regulator, and just look at us — it sounded like a joke. But you keep banging your head against the wall, and then four years later — I mean, this took four years. We were so deeply committed: we're not going to launch, we don't care. We had competitors that launched and did stuff, but until we made it legal... And then we actually did make it legal, and after that I think the company really skyrocketed.

I think in some ways it just gave us the right to go really big. Because I always thought about it — the more I think about what we're really doing at Kalshi, we're really expanding the universe of financial markets to things of relevance that the average person can understand, participate in, or care about. And we could see it. We were sort of not there for a while, and then when we did actually manage to get regulated and launch, and appeal to institutions, appeal to the large distribution like brokers and others — you can think of this as the next generation New York Stock Exchange, because it is, and it's starting to feel like it. It trades like it. The participation, at least from individuals, is much broader and bigger. We had to do it clean and regulated, because otherwise people wouldn't trust it.

So Ari is a smart investor, is the way I'd put it. But basically, right now we're like a hundred billion dollars in transactions a year, and it's growing something like 50% month over month. The more important thing is the market is growing at such a pace that we don't even think much about competition. We just think about, okay, how do we keep growing this market, and we just have to stay market leader. And you're seeing a lot of companies, whether it's financial brokers like Robinhood and Schwab, get in — we're working on some stuff on that front. There's a shift in consumer behavior. I think we've always focused on the robust fundamentals. I really think people see and hear about Kalshi — it's been a year and a half, but I always say this:

this is an overnight success eight years in the making. There's six years of foundation building that I've always focused on. And the thing I like about the company is nothing's really changed in how we operate from those years to today. It's kind of like — we were traders, so we focus on the output, not the outcome. The outcome you can't control. Sometimes you get to the World Cup final and you get to penalties, and you win, and sometimes you lose. And Messi now is the greatest of all time because he won. But if that penalty had gone out, he would have lost. And you can't control the variance, but you have to just make sure that you're consistently putting out output. And that's how we think about it as a company.

I'm still figuring out me as a founder and CEO. That's an interesting question, because when the business scales this fast, the difficult thing about it is you basically have to change yourself every three to four weeks. By the time you start feeling like, okay, you got a handle on this next phase, you're now already like, oh shit, this is not fitting anymore. So that's difficult. And my general answer to this is just work as hard as humanly possible, deliver as much as you can, and generally stay true to your principles.

This year we're battling a lot with, once we're getting a lot of visibility. So people talk a lot about the company, and there's a lot of skeptics and there's a lot of supporters, and you have to start thinking very critically about, okay, what is the role that we want to play, and how do we want to present ourselves.

But I think it just goes back to the vision. I think at the end of the day, long-term success — putting aside commercial success here — is, I think there's a very big opportunity to redefine the way that people ingest news and think about the future. We can really depolarize and objectify, make the whole process more objective, where people are not necessarily listening to pundits. I mean, already we've seen this trend of people don't really trust authority figures anymore. Polls — people are generally skeptical of polls, increasingly so. And I think that's opening up room for an alternative. And I think prediction markets are the antidote. And I think the beauty here is, at the time people would think maybe social media was going to be the antidote, but the issue with social media is in some ways it incentivizes clickbait, whereas prediction markets — the way that we make money at the core is we incentivize truth. The more liquidity, the more participation we get, the more accurate these markets get. So my incentive is actually, how do we make these more and more effective, which means more and more truthful. And I think if we delivered on that, this would be a very, very valuable and useful thing for society.

Which competition?

Well, they don't make money.

Well, there's your difference. But yeah — one thing that we get associated with often times, or some of the skeptics maybe call prediction markets another form of betting. And I always thought — and there's a lot of history about that, and it's interesting — when I first started digging into this and I was like, okay, why does this not exist? And the reason I got excited about maybe we can make it happen is because the journey of all the traditional financial markets was very similar, and there's a lot of parallels. So at the end of the 19th century there was actually a Supreme Court decision, or a big legal battle between the states and the federal government, about whether grain futures or commodity futures should be regulated or considered to be gambling versus a financial instrument. And the critics were saying, well, it should be gambling, because some people are speculating — you know, the age-old question, is the stock market speculation, and so on. And the Supreme Court in 1905 decided that actually, yes, some people are speculating, which you could think of as a bet, but the pricing that comes out of these markets, or the hedging ability for the farmers, was very useful. So that's a financial market.

So we're going to keep it in the financial market bucket. And I think Kalshi has sort of brought back that debate over the last half a decade. And I think it goes back to, one, is there economic and social relevance to the thing outside of just rolling a dice and betting on it? And I think the answer is unequivocally yes here. And then two, I think the mechanic — I think the beauty and the reason why prediction markets are growing so fast is there's no house. You're not going and betting against the house, where the house's revenue, the way they make money, is your losses. A company where the revenue is equal to your losses is going to have a different set of incentives than a company where people are trading against each other. Rich is trading against Ari, Ari is trading against Rich, and I take a transaction fee on top. And my incentive is keeping you trading and more engaged and more interested, making sure you understand what you're doing. I don't care who wins or loses. I'm neutral. And so it enables me to be more transparent and more fair.

When Uber showed up, the taxis weren't saying, 'Hey, we're being disrupted by Uber.' They were saying Ubers are not safe. That's usually the argument. And the same thing happened with Airbnb. You would take one example of one weird host and blow it up, and it's the same thing. It's the age-old story of, there are entrenched interests that are at least perceiving... I don't actually believe necessarily that this is going to be a net loss for anybody. I think everybody's going to be a net winner, but there's some degree of disruption. And when there's those types of dynamics — I mean, Jeremy, you dealt with it a lot in your business too — the entrenched incumbents are going to come out and say all sorts of things. I mean, they generally make up data. I mean, all sorts of anything, you know.

But the truth is, we're a company that spent four years getting regulated up front. We actually are more regulated and have more protections around customer protection and market integrity than really any of these businesses. And if anything, our incentives as a business are healthier, right? Because we just discussed — the reason you have these issues around customer protection and problem gambling is because you've kind of created them, right? If you make money on customers' losses, what are you going to do? You're going to promote more losses, right? And then you're going to have to navigate around the edges about how to mitigate those losses. We don't have that intrinsic incentive. And yes, there's risks with any financial markets. There's excessive risk-taking in the stock market. There's excessive risk-taking in traditional betting markets, and there is in prediction markets. And it's all about having the right set of safeguards. You have to take it seriously. You have to make sure if anybody's doing too much, they have the tools to basically block themselves, or you can block them yourself, and we have all of these things.

I think competition is generally pretty much always consistently a good thing for a marketplace. I think an industry doesn't become an industry by one player. It has to have competition; it has to have a healthy degree of competition. I think it's just so early. It's a bit like AI right now: how much cannibalization is coming from Anthropic versus OpenAI, and they're all still growing, right? And we saw it in crypto for a while. So I think now it's all about growing the market and growing the pie and educating people what this is. And so we actually want more competition. I think they'll do more than they did last year, and we'll do more than we did last year, and I think everybody's a net winner.

So half of the markets come from our team. We have a team that basically is figuring out at all times what is trending, what is top of people's minds. And the second half is our customers suggest things, and then there's a whole setup — and this is actually pretty hard. It has to go through a whole pipeline of how do these become legal financial markets, because each one of these is a market. It's a bit like a stock: it has its own ticker, it's reported, all of that. And there's challenges with that. The two key challenges: one is, how do you define what actually happened? And one of the most complicated markets we've ever had to define, surprisingly, was the election. And most people don't know — the election is not final on election day. I guess most people know that, but most people don't know that the election is not final after Congress certifies the election. So that is not when you know who's going to be the next president. The only way to know who's going to be the next president is for that president to be sworn in on inauguration day. And all these details you have to deal with, because you cannot have uncertainty in markets.

And so that's a whole piece of it. And then there's what we don't do. So it's illegal to do wartime, assassination, violence, but also any markets that... you may have heard these, but these are usually done on unregulated markets. They're not done on regulated markets in the US. But we generally think very hard about making sure no one can profit from a bad thing happening. There are categories of markets that maybe are not strictly illegal but could still be bad, like wildfires. Yes, it would be good to know if there's going to be a wildfire, but you could incentivize someone doing something bad. Yeah. So we don't do these types of markets, and it's hard. I mean, the first time we listed markets it took us 18 months. Now we can do it in 24 hours. So just day over day we improved.