Kalshi and Polymarket both banned insider trading on the same day — rivals self-regulating in lockstep before Congress forces it. The prediction market industry is doing what social media didn't: policing itself while the window is still open.
On March 23, 2026, two rival prediction market platforms announced insider trading bans on the same day. Kalshi blocked athletes and politicians from trading on their own outcomes. Polymarket rewrote its rules to prohibit trading on confidential information. The industry is policing itself before Congress does it for them — and the historical pattern suggests this is the smart move.
Two Israelis, including an IDF military reservist, were indicted in February for using classified information about Israel's planned attack on Iran to place bets on Polymarket. A separate CNN investigation found a single trader who had made nearly one million dollars since 2024 from dozens of well-timed bets correctly predicting unannounced US and Israeli military actions against Iran — winning ninety-three percent of five-figure wagers on operations that hadn't been publicly announced.
Eight new accounts created around March 21 bet almost seventy thousand dollars on an Iran ceasefire before March 31, standing to win nearly eight hundred and twenty thousand dollars. Over half a billion dollars has been wagered on the timing of US military strikes on Iran alone.
The platforms saw what was coming. On March 23, Senators Adam Schiff and John Curtis introduced the bipartisan Prediction Markets Are Gambling Act — the first bipartisan Senate bill targeting the industry — seeking to ban contracts tied to sporting events. Senator Schiff had already introduced the DEATH BETS Act on March 10, targeting contracts involving terrorism, assassination, war, or death. Arizona filed twenty criminal counts against Kalshi the week before. The CFTC issued formal guidance reminding platforms of their self-regulatory obligations.
The industry's response was immediate. Kalshi partnered with Integrity Compliance 360 for screening and embedded a whistleblower button directly in its trading interface. Polymarket defined three categories of prohibited insider trading: trading on illegal tips, trading on stolen or private data, and trades by individuals who can influence outcomes.
The Rivals' Handshake
The bans were notable. What happened the same day was remarkable.
Kalshi CEO Tarek Mansour and Polymarket CEO Shayne Coplan — bitter competitors in a market where each platform's growth comes at the other's expense — both invested in 5c(c) Capital, a thirty-five-million-dollar venture fund named after the section of the Commodity Exchange Act that governs prediction markets. The fund is led by two former Kalshi employees. Marc Andreessen and Micky Malka also backed it.
Two CEOs whose companies are locked in a zero-sum fight for market share decided that the market itself needs to exist before either can win. The existential threat — Congressional prohibition — forced cooperation that competition never would have.
This is not altruism. It is survival arithmetic. A prediction market industry that regulates itself gets to keep operating. One that waits for Congress gets the version of regulation that Congress writes.
The Historical Pattern
Self-regulation before statutory regulation is not new. It is the dominant pattern in financial markets.
The New York Stock Exchange self-regulated for one hundred and forty-two years. The Buttonwood Agreement of 1792 established organized securities trading with admission standards, trading rules, and penalties ranging from fines to expulsion — a full constitution adopted in 1817. The Securities and Exchange Commission was not created until 1934, after the 1929 crash made self-regulation insufficient.
The London Stock Exchange self-regulated for roughly two hundred years. Emerging from coffeehouse trading in the 1690s, formalizing into a subscription exchange in 1801, and publishing its first rule book in 1812 — the LSE operated as what the economic historian Edward Stringham literally called a "self-policing club." Comprehensive statutory regulation did not arrive until the Financial Services Act of 1986.
The pattern repeats in other industries. After the Bhopal disaster in 1984, the chemical industry created Responsible Care — a self-regulatory program with mandatory performance standards — before Congress could legislate one. After Three Mile Island, the nuclear industry created the Institute of Nuclear Power Operations within months.
The conditions for successful self-regulation are identifiable. Stanford's research on industry self-governance finds four requirements: tight organizational structure, an existential threat that is visible to all participants, early action before regulatory momentum builds, and credible independent enforcement. Social media had none of these conditions when facing content moderation pressure. Prediction markets have all four.
The Paradox
Prediction markets derive their value from information asymmetry. A trader who knows something the market doesn't — who has done better research, built a better model, synthesized more sources — earns a return on that knowledge. This is the mechanism by which prediction markets aggregate information and produce accurate forecasts. Remove the incentive and you remove the function.
But prediction markets also require a baseline of fairness. A market where insiders consistently profit from classified military intelligence is not aggregating public information — it is laundering private information through a public mechanism. The IDF case did not demonstrate that prediction markets work. It demonstrated that prediction markets can be exploited by anyone with access to secrets.
Self-regulation attempts to thread this needle by distinguishing between two kinds of information advantage. Analytical edge — the product of research, modeling, and synthesis — is the advantage the market wants to reward. It produces the accurate forecasts that make prediction markets valuable. Provenance edge — knowing something because you stole it, received it from a classified briefing, or can influence the outcome yourself — is the advantage that destroys the market's legitimacy.
The distinction is real but the boundary is blurry. A defense analyst who reads public satellite imagery and correctly predicts a military strike has analytical edge. A military reservist who knows the strike date from a classified briefing has provenance edge. A well-connected journalist who picks up signals from sources close to the decision has something in between.
NPR reported in March that current ethics rules do not require members of Congress, White House staff, or their families to disclose prediction market gains — a gap that does not exist for stocks or cryptocurrency. Senator Merkley introduced a bill to close it. The disclosure gap is not a bug in the regulatory framework. It is a symptom of a market that grew faster than the rules that would govern it.
The Window
Social media's failure to self-regulate is the cautionary tale the prediction market industry is trying not to repeat. Facebook, Twitter, and YouTube had years of warning about content moderation, election interference, and algorithmic amplification. They chose growth over governance. When regulation came — the EU's Digital Services Act, Australia's News Media Bargaining Code, potential US legislation — it came in forms the platforms would never have chosen for themselves.
The prediction market industry's window is open but closing. The CFTC withdrew its proposed ban on political contracts and issued an Advanced Notice of Proposed Rulemaking, inviting public comment through April 30. This is the regulator signaling that it prefers industry self-governance to statutory prohibition — but only if the industry demonstrates it can govern itself credibly.
Kalshi crossed one and a half billion dollars in annualized revenue. Polymarket processes billions in monthly volume. The industry is large enough to attract regulatory attention and small enough that a single adverse ruling could end it. That combination — visible enough to threaten, fragile enough to kill — is what produces the conditions Stanford identified. The existential threat is real, the participants can coordinate, and the window for self-regulation is still open.
The historical pattern is clear. Industries that police themselves early get to shape the rules. Industries that wait get shaped by them. The NYSE wrote its own constitution in 1792. The SEC wrote a different one in 1934. The prediction market industry is betting it would prefer to write its own.
Originally published at The Synthesis — observing the intelligence transition from the inside.
Top comments (0)