Defense Secretary Pete Hegseth warned Pentagon personnel that betting on military outcomes can disclose operational plans and compromise professional judgment. Jon Brodkin reports on an October 20 memo covering prediction markets such as Kalshi and Polymarket alongside traditional betting services. It prohibits financial transactions informed by nonpublic information, whether classified or unclassified, including bets placed through someone else.
The memo describes a leak that can happen through the wager itself. A person with advance knowledge buys a position on an operation’s timing or outcome. Foreign intelligence services can combine that trade with other signals to infer movements, targets or dates. The person collecting the profit and the people exposed by the information are different groups.
Hegseth also identifies a conflict before any information reaches an outside observer: “A logistician with money riding on a strike date has an interest in that date.” An analyst betting on an outcome acquires a financial reason to favor it. Selling an event contract, a bet on whether something happens, creates particular risks when customers can influence the event they are betting on.
Brodkin places the warning alongside earlier suspicious trading. An Army soldier was charged in April with using confidential information to win $410,000 on Polymarket bets about the capture of Nicolás Maduro. An August report by the Anti-Corruption Data Collective identified 152 suspicious wallets that won a combined $8 million on military and defense markets. Suspicious trading patterns alone do not establish that each account used inside information.
Kalshi and Polymarket both prohibit insider trading, according to the article. The memo adds the prospect of administrative, disciplinary and criminal consequences under military or federal law. Its significance is the harm it describes beyond an unfair payout: an individual wager can expose information or introduce a private financial interest into a public decision. Brodkin’s report documents the warning and earlier cases; it does not measure whether the memo changes behavior.