Prediction Markets Promised the Wisdom of Crowds. They're Delivering the Wisdom of Insiders.

The people winning big on prediction markets aren't smarter than you. Some of them just already know the answer. A Bloomberg investigation found that at least 34,000 trades on Polymarket between August 2025 and June 2026 were flagged as potential insider trades — bets placed by accounts that behaved less like forecasters and more like people reading tomorrow's newspaper today.
The detail that lingers: one trader flew to San Francisco for Super Bowl LX and put more than $50,000 on the national anthem running under 117 seconds. Rehearsal information, in other words, priced to the second. Prediction markets were sold as machines for aggregating dispersed public knowledge. They are proving equally efficient at monetizing concentrated private knowledge.
The insider is the product working as designed
Here's the uncomfortable mechanism underneath the scandal. A prediction market pays most to the trader with the best information. Its entire epistemic value — the reason economists defend it — is that it drags private knowledge into a public price.
So what do we call the wedding planner betting on a celebrity engagement announcement, or the staffer trading the timing of military strikes? In equities, that's a crime with statutes and case law behind it. In event contracts, it's often just... the market being informative. The platforms are racing to build surveillance for a new class of insider trader that existing law never contemplated, because the events themselves — anthem lengths, album drops, airstrikes — have no issuer, no fiduciary, and no disclosure regime.
The platforms' defenders make a fair point, and it deserves stating plainly: most flagged trades are probably sharp analysis, not leaks. Pattern-matching a confident bettor to an insider is exactly the error securities regulators spent decades learning to avoid, and 34,000 flags is not 34,000 crimes. Some fraction of those accounts simply did the work — counted rehearsal footage frames, read shipping manifests, modeled turnout. That's the wisdom of crowds functioning.
But the defense concedes the core problem. Nobody, including the platforms, can currently tell the analyst from the leaker. In equities, that distinction is the entire regulatory apparatus. Here it's a vibe.
You can't have it both ways forever. Either the market's edge in truth-finding comes from insiders trading their knowledge into the price, or the platforms scrub insiders out and the prices become polls with fees.
Meanwhile, the money got serious
None of this ambiguity has slowed the buildout. Jump Trading doubled its prediction markets team to ride record volumes. Venture money keeps arriving — Fortune reported a $9 million raise for Pascal, a startup built to challenge Kalshi and Polymarket head-on. Midterm election contracts alone have already cleared $197 million in volume across 1,408 open markets, by NBC's count, with three-plus months still to run.
And the lobbying arms race tells you everyone expects the fight to be decided in Washington, not in court. Kalshi spent $500,000 lobbying federal policymakers in the second quarter and now retains seven firms, while the casino industry funds the opposition. When both sides are paying that much to define what a "bet" is, the answer stopped being a legal question and became a purchasing decision.
The states already lost round one
Arizona tried the direct approach. In March, its attorney general filed the first criminal charges in US history against a major prediction market, accusing Kalshi of running an unlicensed gambling operation, complete with election wagering. The CFTC responded by suing Arizona, and a federal judge first blocked the prosecution at the agency's request and then made the block permanent, ruling the state's case unconstitutional.
Read that sequence again. A federal regulator went to court to stop a state from criminally prosecuting a company the regulator supervises. Ohio, New York, Wisconsin and Connecticut are running variations of the same fight. Federal preemption is winning, which means the surveillance question — who catches the wedding planner, the rehearsal staffer, the aide with the strike timing — lands entirely on the CFTC and the platforms themselves. The same platforms whose revenue scales with volume, including the informed kind.
Crypto should recognize this movie
There's a reason this all feels familiar to anyone who lived through DeFi's adolescence. Polymarket runs on crypto rails; its markets settle in stablecoins on-chain, and every flagged trade in Bloomberg's dataset is publicly auditable in a way Wall Street's dark pools never were. The transparency that makes the insider problem visible is itself a crypto export. Traders tracking these markets alongside their portfolios in apps like The Crypto App can watch the odds move in real time — which also means they can watch the odds move before the news, which is precisely the problem.
Crypto learned this lesson the expensive way: transparency without consequences just documents the crime beautifully. On-chain forensics made rug pulls visible; it took enforcement, and better market structure, to make them rarer.
Prediction markets are now where token markets were in 2021 — explosive growth, genuine innovation, and a regulatory perimeter drawn in pencil. The volumes will keep breaking records through the midterms. The real test arrives the morning a market correctly prices an unannounced merger, a Fed decision, or a casualty count before the public knows, and everyone can see the wallet that knew first.
What happens then decides whether prediction markets become financial infrastructure or a very well-instrumented crime scene.