Defense Secretary Pete Hegseth has ordered the Pentagon’s workforce to stop betting on its own operations. A memo dated Oct. 20 bars military personnel, intelligence professionals and department civilians from using nonpublic government information to trade on prediction markets, singling out Polymarket and Kalshi by name, with violators facing discipline under the Uniform Code of Military Justice and, where warranted, federal prosecution.
I have spent most of this year arguing that this moment would arrive.
In January I wrote in War on the Rocks that markets tied to military outcomes structurally reward insider trading and hand participants a financial interest in how wars unfold.
In February, Max Meizlish and I argued in the Wall Street Journal that gambling markets had become an intelligence asset that adversaries would read whether or not Washington chose to, and in May I described in the Journal how cheating on these platforms was evolving from simple foreknowledge toward the manipulation of outcomes themselves.
Each of those claims drew polite doubt from people who still regarded prediction markets as a novelty, and the Pentagon has now written the premise into official policy.
The prohibition was overdue by any measure. In April the Justice Department charged an Army Special Operations master sergeant with earning roughly $410,000 by betting on the capture of Nicolás Maduro before the raid became public, and over the summer researchers identified more than 150 wallets that appear to have traded on military insider knowledge while Polymarket referred dozens of cases to prosecutors. The department could hardly go on pretending the problem was hypothetical.
Yet the memo’s lasting significance lies in a single sentence of justification from Mr Hegseth:
“Our enemies watch movements in these markets,”
With that line the Secretary of Defense conceded what the department’s personnel policies had never addressed, which is that prediction markets have become a collection target and the United States sits on the wrong side of the trade.
Consider what a hostile intelligence service sees.
A contract on an American strike or a leadership decapitation idles at a low probability for weeks, and then, in the hours before an operation, the price lurches upward on heavy volume. The market has aggregated the knowledge of whoever placed those bets, stamped it with a time and published it to the world, so that no recruited asset, intercept or satellite pass is required because the leak arrives already priced and sorted.
The memo also reaches a subtler danger, which Mr. Hegseth frames in operational terms. A logistician with money riding on a strike date, he observes, acquires an interest in that date, and an analyst holding a position on an outcome acquires an interest in that outcome, a condition he describes as a dilution of the warrior ethos.
I have been calling it the agency problem, because a wager converts a spectator into a participant with a reason to influence the result and allows the people who serve a principal to profit by betting against him. In a trading firm that is a compliance matter, while in a combatant command it is a matter of lives.
Here’s what the memo leaves unsaid. If adversaries are harvesting intelligence from these markets, then the United States can do the same.
Open-source intelligence formalized the systematic reading of public information, social media intelligence did the same for platforms, and a parallel discipline is now taking shape around markets where people stake money on outcomes, a field practitioners have started calling betting intelligence, or BETINT.
Its questions are the familiar ones of any collection discipline: who is moving a given market and when, whether the timing of a position implies foreknowledge of an event, and which contracts adversaries have adopted as reference points. None of it requires classified access, only tradecraft applied to a data source that did not exist at scale three years ago.
The need will only grow. Prediction markets are on course to become a multitrillion-dollar asset class, and contracts touching military, diplomatic and public-safety outcomes will multiply along with them. Forbidding soldiers from participating removes one category of bettor without shrinking the signal that remains or blinding the foreign services already reading it.
Mr. Hegseth has told his department that the enemy watches these markets. The harder question, which his memo raises without answering, is who in the American intelligence community is watching back.
BETINT stands ready to assist that mission.
Disclosure: AI tools assisted with drafting and editing this piece; the analysis, arguments and conclusions are my own.
