Cart
0
150+ Polymarket Wallets May Have Traded on U.S. Military Secrets
More than 150 wallets on Polymarket International have been identified by researchers as potentially trading on non-public U.S. military information. The wallets reportedly made about $8 million with an extraordinary 97.2% average win rate, raising questions about insider trading, prediction-market oversight and whether public betting activity could expose military signals to foreign adversaries.
Polymarket company logo from https://commons.wikimedia.org/
A prediction market is supposed to turn uncertainty into a wager.
But researchers now say some traders may have been betting on information that ordinary investors simply did not have.
Gear Spotlight- What Our Readers Are Picking Up
On August 20, 2026, the nonprofit Anti-Corruption Data Collective, or ACDC, published research identifying more than 150 Polymarket wallets that it believes may have traded using inside information connected to U.S. military activity.
The numbers are what immediately drew attention.
ACDC identified 152 military and defense-related wallets that collectively made approximately $8 million, with an average win rate of 97.2%.
That does not prove that all 152 wallets belonged to government insiders.
ACDC itself acknowledges that unusual success can have other explanations, including luck, and that not every person who trades on military markets would fit the group's methodology.
The researchers began with a much larger group.
They identified 556 wallets, which they called “Orcas,” after traders who placed unusually large, highly successful long-shot bets in specialized markets where an insider could potentially possess an informational advantage.
The methodology focused on bets of at least $2,500, accumulated within an hour or less, on outcomes carrying odds of 35% or less.
From that broader group, the researchers isolated the 152 wallets involved in military and defense markets.
The concern is not simply that someone may have made money.
The bigger concern is what happens when those bets become visible to everyone else.
Polymarket International uses blockchain technology, meaning transactions are publicly visible even though the identities behind individual wallets can remain anonymous.
That creates an unusual intelligence problem.
A person with privileged information might place a bet anonymously, but the bet itself can become a signal.
ACDC says larger “Whale” traders and automated bots appeared to copy some of the suspicious trades, potentially amplifying the original signal across the market.
In a military context, that could become more serious.
If a trader knows something about an upcoming operation before it becomes public, a prediction-market wager could reveal that information indirectly.
Other traders might copy the position.
Algorithms might detect the unusual activity.
And, potentially, foreign intelligence services could monitor the blockchain for the same patterns.
The issue has already moved beyond theory.
In April 2026, U.S. soldier Gannon Ken Van Dyke was charged with allegedly using classified information to bet on the expected removal of Venezuelan President Nicolás Maduro. Prosecutors say he made approximately $400,000 from the bets. Van Dyke has pleaded not guilty.
Interestingly, ACDC says Van Dyke was not included among its 152 military “Orcas” because his trading pattern developed more slowly.
That distinction matters because it shows the research is identifying a particular behavioral pattern, not claiming to have identified every suspected insider.
Polymarket has previously said it monitors suspicious activity and has referred dozens of trader wallets to authorities, including in the Maduro case. The company did not respond to Reuters' requests for comment on the latest research.
The Pentagon also declined to comment specifically on the findings, saying it does not comment on intelligence-related matters or third-party research.
The debate now extends into regulation.
ACDC argues that prediction platforms should strengthen identity verification and consider restricting markets where access to non-public information could provide an especially valuable advantage.
The challenge is balancing transparency with security.
Prediction markets can provide useful information about public expectations, but military operations are different from elections, sports or economic indicators.
A bet concerning an upcoming military operation could potentially reveal something about an operation before officials publicly acknowledge it.
And there is another complication.
Even if the original insider is never identified, the market itself can potentially distribute the signal.
That means the question facing regulators and defense officials is no longer simply whether someone illegally profited from classified information.
It is whether prediction markets could unintentionally become another channel through which sensitive military information travels.
For U.S. troops and commanders, that distinction could become increasingly important as prediction markets grow and military activity becomes more heavily tracked through public data.
ACDC's research does not establish that the 152 wallets were operated by military personnel, or that classified information was used in every case.
But a 97.2% average win rate across hundreds of millions of dollars in military-related market activity is unusual enough to attract scrutiny.
The next question is therefore much harder to answer.
If some traders really did know what was coming before everyone else, how much did their bets reveal before the public knew it was happening?
Editor's Note: The research identifies wallets as potentially suspicious based on trading patterns. It does not publicly establish the identities of the wallet owners or prove that all 152 traders used classified information. The investigation and regulatory debate remain ongoing.
