Two prediction market startups announced they're winding down within 90 minutes of each other on Monday morning, giving users until Sept. 30 to withdraw their funds. The closures hit opposite ends of the prediction market spectrum, and they leave the field almost entirely to Kalshi and Polymarket, which together control 93% of volume.
A Monday morning double exit
The two announcements came back-to-back, a timing that suggests coordination or at least a shared sense of urgency. Neither startup gave a reason in the public notices, but both set the same withdrawal deadline. Users have roughly seven weeks to move their money out. The speed of the double exit is unusual — most shutdowns come with more lead time, and two in one morning is a rare sight.
Trepa's Solana-based approach
Trepa built its own mechanism on Solana, paying users based on participation rather than just the outcome of a bet. That design put it at one end of the prediction market spectrum, where the other venue apparently took a more conventional route. The shutdown leaves that niche empty. Trepa's model was a bet on engagement, not just accuracy, and it's unclear if anyone will pick up that approach.
A market now dominated by two
Kalshi and Polymarket hold 93% of prediction market volume, according to the latest figures. That leaves little room for smaller players, and Monday's exits suggest the squeeze is real. The two startups were at opposite ends of the spectrum, but both found the market too tough. With these two gone, the remaining 7% of volume is spread across a handful of tiny venues, none of which have shown signs of breaking out. The duopoly isn't new, but it's now more entrenched than ever.
The Sept. 30 deadline
Both startups told users to withdraw by Sept. 30. That's a hard date, and anyone with funds in either venue should move them before then. After that, the platforms will be gone. Users who miss the deadline will have to chase whatever process the startups set up post-shutdown, which is rarely a pleasant experience. The clock is ticking.
