SEON CEO: Prediction markets can forecast the future. Can they survive their own manipulation problem?

Prediction markets have moved from niche internet experiments to one of the fastest-growing segments in financial services, with projections pointing toward trillions in annual volume. Robinhood’s CEO recently declared that we’re in a “prediction markets supercycle.” The vocabulary has shifted with the market. Gambling gave way to investing in outcomes. Bets became event contracts. Your edge, your return. The rebranding isn’t accidental, and it’s not entirely wrong. Prediction markets genuinely aggregate real information, and when real money is on the line, people don’t lie—. CNN’s partnership with Kalshi proved the point during the 2024 elections: crowd-sourced probability estimates can outperform traditional polling throughout the election cycle, precisely because participants have something real at stake: money. But going mainstream doesn’t solve the problem these markets have been quietly carrying. It amplifies it. The Difference Between Predicting an Outcome and Engineering One The fundamental premise of prediction markets is that outcomes can be observed, not influenced. Once that assumption breaks, you don’t just have a fraud problem. You have a structural collapse of the product’s entire value proposition. Consider a scenario that has made its way through compliance circles: a presidential speechwriter bets heavily on whether a specific, obscure word will appear in a major address, then ensures it does. No money changes hands illegally, and no law is obviously broken, but the market has been gamed without any honest participant knowing. The Commodity Futures Trading Commission’s (CFTC) own Bloomberg Law analysis confirms that applying traditional insider trading rules, which generally require a trade made while in possession of material non-public information in breach of a legal duty, isn’t always straightforward in prediction markets. This vulnerability isn’t hypothetical a