BlockBeats News, August 21st, a recent study by the Anti-Corruption Data Collective (ACDC) revealed that at least 152 anonymous wallets on Polymarket may have traded based on advance knowledge of U.S. military and defense information, accumulating profits of around $8 million with an average success rate of 97.2%. However, the research institution emphasized that blockchain data itself cannot prove the actual identity of these wallets' controllers or confirm their information sources, therefore, insider trading cannot be concluded based on this.

ACDC analyzed settled markets on Polymarket, focusing on identifying "low-probability high-stake" behavior, defined as transactions where at least $2500 was collectively bet within 1 hour on an outcome with a probability of no more than 35%. The study identified 556 wallets exhibiting abnormal trading patterns, with 152 involving military and defense markets. These wallets were referred to by researchers as "Orcas," characterized by sudden appearance, betting on low-probability events, achieving exceptionally high success rates, and subsequently exiting with profits.

The study found that the success rate of low-probability bets in military and defense markets was significantly higher than the overall Polymarket level, with some abnormal trades quickly attracting large traders and automated trading bots to follow suit.

For example, prior to the U.S. attack on an Iranian nuclear facility in June 2025, after an unusual bet on a U.S. military action, an automated trading bot subsequently wagered $200,000, and another large trader bet $100,000. Researchers also discovered similar abnormal betting and follow-up trading behavior before the U.S. struck Tehran.

David Szakonyi, co-founder of ACDC, stated that unusual trading activities on Polymarket are more publicly visible than many traders realize, with large traders and bots already tracking and mimicking suspected insider trades. Therefore, foreign intelligence agencies may also monitor these publicly available on-chain trading activities.

Recently, insider trading in prediction markets has received ongoing regulatory attention. In April, the U.S. Commodity Futures Trading Commission (CFTC) charged U.S. Army personnel Gannon Ken Van Dyke with trading Polymarket contracts using confidential information about the capture of former Venezuelan President Maduro, allegedly profiting over $404,000. This case is not directly related to the 152 wallets identified in this study.

Furthermore, in May of this year, the CFTC also charged a Google software engineer with trading 23 Polymarket contracts based on confidential information related to the company's 2025 "Yearly Search" rankings, earning approximately $1.2 million with nearly perfect accuracy.

Polymarket had previously stated that the company would closely monitor suspicious transactions and had reported dozens of wallets to relevant authorities. As the size of the prediction market continues to grow, insider trading, market manipulation, and national security risks are becoming key reasons for regulatory authorities to enhance oversight.