More than 150 wallets on Polymarket International may have traded on inside U.S. military information, attracting copycat bets and raising concerns that the market could broadcast signals exploitable by foreign adversaries, according to research published on Thursday.
The analysis by the nonprofit research group Anti-Corruption Data Collective, first reported by Reuters, sheds new light on likely military insider trading on the prediction market and the broader risks such activity may pose.
It comes as policymakers increase scrutiny of the booming industry following bets ahead of U.S. military actions, including that of a U.S. soldier charged in April with using classified information to bet on the removal in January of Venezuela President Nicolas Maduro, which lawmakers have said could hurt national security.
Spokespeople for Polymarket, which was founded in 2020, did not respond to requests for comment. Polymarket says it has strict controls, closely monitors for suspicious activity and has referred dozens of trader wallets to the authorities, including the Maduro case.
Polymarket's international platform settles trades on a blockchain, meaning wagers are public, although traders remain anonymous. ACDC analyzed all settled markets through May 5 to identify "long-shot" bets, which it defines as wagers of at least $2,500 placed cumulatively in an hour or less on an outcome with odds of 35% or less.
Within that universe, ACDC found 556 wallets it dubbed "Orcas," a nod to the killer whale's precise and selective hunting tactics. These traders typically opened an account and quickly placed a highly successful long-shot bet in niche markets where insiders may have an informational edge, before in many cases cashing out and disappearing.
Of those, there were 152 exceptionally successful wallets that bet on military and defense markets, making $8 million collectively, with an average win rate of 97.2%.
Orca traits may have other explanations, including pure luck, ACDC said. Likewise, not all potential insiders are Orcas. Gannon Ken Van Dyke, the soldier who prosecutors say made $400,000 betting on Maduro's ouster, built his position more slowly and is therefore not among the 152 military Orcas identified. Van Dyke has pleaded not guilty.
Still, those wallets are the most likely to be trading on military secrets, according to the report. While some of those 152 wallets have previously been flagged by other researchers and the media, ACDC says it has discovered dozens more previously unreported wallets.
The Department of Defense does not comment on intelligence‑related matters or the findings of third-party research, a spokesperson said.
COPYCATS MAY AMPLIFY INSIDER SIGNALS
Profiting from confidential government information is generally illegal, but of greater concern is evidence that military Orca trades appear to attract copycat bets from deep-pocketed "Whales" and automated trading bots, ACDC said.
Although copycat trading itself is legal and has previously been documented on prediction markets, ACDC found signs that it may amplify Orca signals and broadcast them more widely.
"Most people vastly underestimate how observable unusual betting activity actually is on Polymarket. It's all right there on the internet, and we can see clear signs that big traders and bots are copying potential insider trades," ACDC co-founder David Szakonyi said. "It would be naive to think foreign-intelligence agencies aren't monitoring these markets."
Polymarket has said its public ledger allows for greater scrutiny. A spokesperson for the Commodity Futures Trading Commission, which is pushing for jurisdiction over prediction markets, did not comment on the findings. The regulator has said that it will strictly police misconduct in the space, and has brought charges in at least three cases so far.
ACDC comprises academics and investigators with expertise in illicit finance and aims to push anti-corruption reforms.
The group argues that all traders should be required to prove their identity, and payouts on suspicious trades be withheld pending investigation. But to truly mitigate insider-trading risks, markets in which the use of non-public information may be most actionable and profitable should be banned, it argues.
"Limiting the type of people who can bet on prediction markets or relying on law enforcement investigations ... will not be enough," it said.
