Just caught up on some really interesting takes from Kalshi's first research conference in New York, and honestly, it paints a pretty different picture of where prediction markets are headed than what most people think.

Everyone talks about election betting and sports, right? But here's what's actually happening: Kalshi's weekly volume jumped from around $10 million a year ago to $3 billion now—that's a 30x increase. Yeah, sports is doing huge numbers right now, especially with March Madness, but the real story is that sports share of total volume is actually at a historic low. Everything else is growing faster. Entertainment, crypto, politics, culture—these categories have way better user retention than sports.

The institutional side is where this gets really interesting. Goldman Sachs, CNBC, Tradeweb—these aren't retail players. They're using Kalshi data as a benchmark for pricing uncertainty. Think about it: traditional markets have the S&P 500 for stocks, ICE for oil. But for political events, macro outcomes, Supreme Court decisions? There was basically nothing until prediction markets created a real-time price discovery mechanism. Now institutions can actually hedge political risk directly instead of trying to correlate it through other assets.

Tarek Mansour from Kalshi laid out how this institutional adoption is actually happening in phases. Phase One is just data application—like how portfolio managers now check Kalshi odds the same way they check the VIX. Phase Two is the compliance and technical integration work. Phase Three is where real hedging starts and the market gets self-reinforcing. Most institutions are still in Phase One, some in Phase Two, only a few at Phase Three.

One thing blocking faster adoption: current margin requirements are brutal for institutional players. You need to deposit the full contract value upfront—to hedge $100, you post $100. That's fine for retail but kills capital efficiency for hedge funds and banks. Kalshi just got its NFA license and is working with the CFTC on margin trading, which could change the game.

What's wild is that Kalshi itself might not even be publicly traded yet, but the institutional adoption curve suggests this space is normalizing fast. Senior political leaders are now citing Kalshi odds publicly. Party committees use prediction market data as standard reference. Vote Hub is integrating Kalshi data into their models. Two years ago, the best traders on the platform were amateurs. Now they're professionals who've been studying this for over a decade.

The comparison that stuck with me: options markets in the 1970s faced similar concerns about manipulation and regulatory uncertainty. Eventually the infrastructure matured, and now it's just... normal. Nobody thinks twice about it. We're probably looking at the same trajectory here, just compressed. Someone at the conference predicted institutional adoption within five years, maybe sooner.

The real question isn't whether prediction markets work anymore. It's how institutions integrate them into their operations. When people stop asking if they should use these markets and start asking how, that's when you know it's become essential infrastructure.