Researchers Flag $8M Military Bets
PHOTO CAPTION: “Polymarket logo is seen in this illustration taken, August 3, 2026. REUTERS/Dado Ruvic/Illustration”
Researchers have identified 152 unusually successful Polymarket wallets betting on military and defense events, raising concerns that people with access to nonpublic information may be using prediction markets to profit from upcoming military operations.
The Anti-Corruption Data Collective released the findings Aug. 20 after analyzing trading activity on Polymarket, an online prediction market where users can wager on whether future events will occur.
The researchers did not identify the people behind the 152 wallets and did not conclude that all of them are military insiders. Instead, the group said the wallets display characteristics that make them among the strongest candidates for potential insider trading.
The group focused on what it calls “longshot” bets — at least $2,500 wagered in a short period on outcomes priced at a probability of 35% or less. It then looked for highly specialized wallets that concentrated their betting in a small number of markets and topics while winning more than 75% of those longshot wagers.
Researchers ultimately identified 556 wallets fitting their broader “Orca” profile. Of those, 152 exceptionally successful wallets were active in military and defense markets.
Those 152 wallets collectively made about $8 million and had an average win rate of 97.2%, according to the analysis. Task & Purpose reported that the group placed roughly $2 million in military longshot bets.
Half of the identified wallets placed their first longshot wager within two days of opening their Polymarket account, according to the researchers. Many concentrated heavily on unlikely events rather than spreading wagers across large numbers of unrelated markets.
The researchers said unusual trading patterns alone do not prove someone possessed classified or confidential information. Luck and other explanations are possible, and the group acknowledged that not every potential insider would necessarily fit its profile.
The concern extends beyond whether individual traders are illegally profiting.
Polymarket trades are recorded on a public blockchain, meaning other traders can watch unusual bets in real time. Researchers found cases in which the specialized wallets placed wagers before major military events and were then followed by larger traders and automated accounts making similar bets.
The group argues that pattern could create an operational-security risk. If traders can identify potentially informed bets before military action occurs, foreign intelligence services could theoretically monitor the same publicly available information.
In one example cited by Reuters, an identified wallet bet on U.S. military action in Iran hours before June 2025 strikes. Larger accounts and automated traders subsequently placed wagers of $200,000 and $100,000 on similar outcomes. Researchers said comparable activity appeared before other U.S.-Israeli strikes on Tehran.
The Pentagon declined to comment on the researchers' findings, telling Reuters that it does not comment on intelligence-related matters or third-party research. Polymarket did not respond to Reuters' request for comment on the new report.
Polymarket has previously said it monitors suspicious activity and has referred trader wallets to authorities. The company has also argued that the transparency of its public blockchain makes suspicious trading easier to identify.
The issue has already moved beyond theoretical concerns.
Army Master Sgt. Gannon Ken Van Dyke was charged in April with using classified information about a U.S. military operation to capture Nicolás Maduro to place wagers on Polymarket.
Federal prosecutors allege Van Dyke was involved in the planning and execution of Operation Absolute Resolve and used nonpublic information about the mission to place approximately $33,934 in wagers on outcomes involving Venezuela and Maduro. Prosecutors say he ultimately made approximately $409,881 in profit.
Van Dyke has pleaded not guilty. His case is separate from the 152 wallets identified in the new research, and researchers noted that his trading behavior did not fit their specific “Orca” profile.
The Anti-Corruption Data Collective is recommending stronger identity requirements for prediction-market users and restrictions or bans on markets where people with privileged government information could have a significant advantage.
The group argues that policing individual traders after the fact may not address the larger national-security problem because potentially informed wagers can become visible to other traders — and potentially foreign governments — before investigators know who placed them.
(Source: OAF Nation)
