The $21 billion-per-month prediction market industry faces its most serious regulatory challenge since inception. On May 22, 2026, House Oversight Chair James Comer launched a formal investigation into insider trading on Polymarket and Kalshi, sending document requests to both CEOs with a June 5 ...
"Members of Congress can't participate in the predictions market, nor can government employees or people in the president's administration." — Rep. James Comer, Chair, House Oversight and Government Reform Committee
The $21 billion-per-month prediction market industry faces its most serious regulatory challenge since inception. On May 22, 2026, House Oversight Chair James Comer launched a formal investigation into insider trading on Polymarket and Kalshi, sending document requests to both CEOs with a June 5 deadline. The probe was triggered by at least $2.4 million in suspicious profits tied to classified U.S. military operations, a federal indictment of an active-duty Special Forces soldier, and analytical evidence showing 80 accounts with a 98% win rate on military action contracts — a statistical impossibility absent material nonpublic information.
The investigation arrives as the industry's two dominant platforms diverge sharply on regulatory positioning. Kalshi, valued at $22 billion after a $1 billion Series F, holds a CFTC designation and commands 72% of combined weekly volume. Polymarket, valued at $15 billion, remains in regulatory limbo for U.S. users despite acquiring CFTC-registered QCEX for $112 million in 2025. Both face existential questions about whether prediction markets on military, political, and national security events can coexist with democratic governance.
House Oversight and Government Reform Committee Chair James Comer (R-Ky.) sent formal letters to Polymarket CEO Shayne Coplan and Kalshi CEO Tarek Mansour on May 22, 2026. The letters demand documentation by June 5 covering:
The investigation follows Rep. Chris Pappas (D-N.H.) calling for subpoena authority. Comer confirmed subpoenas may follow if voluntary compliance is insufficient. The probe is bipartisan: both Republican and Democratic members have publicly supported the inquiry.
This is not the first regulatory action of 2026. The timeline of escalating enforcement includes:
| Date | Event | Entity | |------|-------|--------| | Feb 25, 2026 | CFTC issues prediction market insider trading advisory | CFTC | | Mar 17, 2026 | Arizona files 20-count criminal information | Kalshi | | Apr 22, 2026 | Three candidates suspended for self-betting | Kalshi | | Apr 23, 2026 | First-ever CFTC insider trading complaint on event contracts | CFTC/DOJ | | Apr 30, 2026 | Democrats urge CFTC to rein in prediction markets | Congress | | May 19, 2026 | Senate unanimously bans member trading | U.S. Senate | | May 22, 2026 | Full House Oversight investigation launched | House Oversight |
On April 23, 2026, the U.S. Attorney's Office for the Southern District of New York unsealed a criminal indictment against Master Sgt. Gannon Ken Van Dyke, 38, stationed at Fort Bragg. The Department of Justice alleged Van Dyke used classified information obtained during his role as a planner of "Operation Absolute Resolve" — the U.S. military capture of former Venezuelan President Nicolás Maduro — to trade on Polymarket.
The mechanics were straightforward:
Van Dyke faces five federal charges: unlawful use of confidential government information for personal gain, theft of nonpublic government information, commodities fraud, wire fraud, and making an unlawful monetary transaction. After the operation succeeded, Van Dyke allegedly attempted to conceal his identity by requesting Polymarket delete his account and changing his cryptocurrency exchange email.
This case represents the CFTC's first-ever insider trading enforcement action involving event contracts. The parallel civil and criminal proceedings signal that both the CFTC and DOJ view prediction market insider trading as a serious offense, not a regulatory gray area.
Blockchain analytics firm Bubblemaps identified a cluster of 80 Polymarket accounts that placed bets on U.S. military actions against Iran with a 98% win rate — including wagers placed hours before undisclosed strikes. The statistical probability of this pattern occurring through legitimate information or luck is effectively zero.
Key findings from the Bubblemaps analysis and subsequent New York Times investigation:
The trades were placed on Polymarket's Polygon-based infrastructure, where pseudonymous accounts can be created with minimal identity verification for non-U.S. users. The on-chain nature of the trades provides immutable evidence but complicates jurisdictional enforcement.
The two dominant platforms represent fundamentally different approaches to prediction market architecture and now face different regulatory exposures.
Kalshi's CFTC designation provides legal certainty and institutional access but also subjects it to direct regulatory oversight. Arizona's 20-count criminal information (March 17, 2026), alleging illegal gambling and election wagering, was temporarily blocked by a federal judge who ruled federal preemption likely applies.
Polymarket's crypto-native infrastructure (Polygon settlement, pseudonymous accounts) makes it the platform of choice for users seeking to avoid identity disclosure — including, allegedly, those trading on classified information. The platform's $112 million QCEX acquisition signals intent to go regulated in the U.S., but the beta rollout has not meaningfully shifted volume from its offshore book.
The regulatory divergence is producing measurable market effects. Kalshi's weekly volume grew from approximately $2.5 billion in January 2026 to $4.13 billion by May 11 — a 65% increase. Polymarket's weekly volume declined from approximately $2.8 billion to $1.60 billion over the same period. Institutional capital appears to be flowing toward the regulated venue as enforcement risk escalates.
Congress has introduced at least five pieces of legislation targeting prediction market abuse since March 2026:
DEATH BETS Act (March 2026) — Rep. Mike Levin and Sen. Adam Schiff. Bans contracts tied to war, terrorism, or assassination.
Public Integrity in Financial Prediction Markets Act (S.4188) — Bans members of Congress and government employees from trading on political prediction markets when they possess or could obtain inside information.
Prediction Market Act of 2026 — Sens. Kirsten Gillibrand (D-N.Y.) and Dave McCormick (R-Pa.). Bans the president, vice president, members of Congress, and senior executive branch officials from trading.
Senate Rule Change (May 19, 2026) — Unanimously adopted rule barring all senators and staff from prediction market trading.
Rep. Ritchie Torres Bill — Specifically targets insider trading on prediction markets in response to the Van Dyke case.
None of these bills ban prediction markets outright. The legislative consensus appears to be: allow the markets to exist, but prohibit government insiders from trading on them and ban contracts on certain sensitive event categories (military operations, assassinations).
The most consequential concern raised during this episode extends beyond individual fraud. National security experts argue that prediction markets create a novel intelligence vulnerability:
Signal leakage: If analytics firms like Bubblemaps can identify suspicious pre-strike trading, so can adversarial intelligence services. According to CoinDesk reporting on May 21, 2026, experts warned that adversaries "can make war plans accordingly" based on market movements preceding undisclosed operations.
Incentive corruption: Military and intelligence personnel now have a financial incentive to leak operational timing. The Van Dyke case demonstrates a 1,108% return in seven days from a $34,000 investment — a payout structure that may exceed what hostile intelligence services offer for equivalent information.
Reverse intelligence: Adversaries may find it cheaper to identify classified information leakers through anonymous betting markets rather than recruiting intelligence assets through traditional means. The pseudonymous nature of crypto-settled markets provides cover.
Market manipulation for disinformation: Large bets placed to move prediction market odds could be used to create false signals about impending military action, potentially triggering real-world responses.
The CFTC's February 2026 advisory acknowledged these risks by requiring exchanges to conduct "real-time monitoring" of suspicious activity. Whether either platform possesses surveillance capabilities comparable to traditional financial market oversight remains unclear.
Prediction markets have scaled from a niche experiment to a $21 billion monthly volume industry in under two years. That growth brought them squarely into contact with the machinery of national security, classified information, and democratic governance. The result is a regulatory reckoning that will determine whether the industry survives in its current form.
The data suggests several outcomes are likely: government employee trading bans will pass in some form (bipartisan support exists in both chambers); sensitive event categories (military operations, assassinations) will face contract restrictions; and platforms will be required to implement surveillance capabilities comparable to traditional derivatives exchanges.
For market participants, the distinction between Kalshi and Polymarket is becoming material. Kalshi's regulated status insulates it from the most severe enforcement scenarios. Polymarket's crypto-native architecture — the feature that enabled pseudonymous trading by Van Dyke and the Iran-linked accounts — is now its primary liability. The $7 billion valuation gap between the two platforms ($22 billion vs. $15 billion) reflects this regulatory risk differential.
The prediction market industry generated real economic value in 2026: price discovery on election outcomes, hedging for policy-sensitive businesses, and information aggregation. Whether that value survives the national security crisis depends entirely on whether platforms can implement insider trading controls without destroying the anonymity and accessibility that drove adoption.