A Casino for Military Intelligence
When insiders bet on classified operations, adversaries are watching
In the early hours of January 3, US special operations forces moved into Venezuela on a mission that depended entirely on secrecy. The objective was to capture the country’s leader, Nicolás Maduro. Details of the operation were so closely held that even senior officials learned about the timing only a few hours before it was underway.
At roughly the same time, a newly created account on the prediction-market platform Polymarket placed tens of thousands of dollars on Maduro’s being out of power by month’s end — a position that stood out for its size, its timing, and its confidence against prevailing odds. When Maduro was captured, the payout was immediate: more than $400,000 in profit.
For a brief moment, it looked like extraordinary luck. According to federal prosecutors, it was anything but. The person allegedly behind the trades was Army Special Forces Master Sergeant Gannon Ken Van Dyke — a soldier involved in planning the operation. He was arraigned yesterday and pleaded not guilty.
Van Dyke’s indictment marks the first case of its kind. It likely won’t be the last. And the question it raises goes well beyond the misconduct of one soldier. It concerns what happens when the mechanisms designed to predict events begin to intersect with the people responsible for carrying them out — as adversaries watch both.
A system built to surface information
Prediction markets operate on a simple premise: people with better information make better predictions when money is on the line. Platforms like Polymarket and Kalshi let users buy and sell contracts tied to real-world outcomes — whether a conflict will escalate, whether a political leader will fall — with prices fluctuating to reflect collective probability estimates.
In many cases, that works as intended. Markets incorporate information quickly, often adjusting faster than official statements or traditional analysis. It’s precisely this capacity that has made them attractive as forecasting tools — a military intelligence bulletin last year even advocated integrating Polymarket data into national security threat assessments.
But these markets don’t distinguish between public information and privileged knowledge. They absorb whatever signals participants bring, regardless of origin. When those signals come from someone with access to classified information, the market doesn’t filter them out. It incorporates them. At that point, what looks like a prediction begins to resemble a leak.
Bets as intelligence leaks
This is where the risk becomes something more than that of financial crime. Prediction markets are public, operate continuously, and are globally accessible. Every price shift, every volume spike, every anomalous position is visible to anyone with the capacity to monitor it. Russia, China, and other powers that maintain sophisticated intelligence operations targeting the US have every incentive to watch these markets systematically and the tools to do it. A loosely regulated, publicly accessible platform trading on the nation’s military secrets is, from their perspective, an open intelligence window that costs nothing to look through.
“The issue isn’t the soldier ‘betting on himself’ — it’s that adversaries can read the order book,” said Alex Goldenberg, a fellow at the Rutgers University Miller Center on Policing and Community Resilience. “Prediction markets on military ops are a real-time intelligence feed for our enemies.”
In the hours before Israeli strikes on Tehran killed Ayatollah Khamenei, Polymarket lit up. Six accounts, funded through cryptocurrency wallets that had been dormant until shortly before, made more than $1.2 million betting on an imminent US strike. Whether the Iranian security apparatus was monitoring Polymarket is unknown. But the signal was there.
Former Air Force Secretary Frank Kendall put the risk of exploitation by adversaries plainly: “All it takes is monitoring for patterns that deviate from established norms. Modern AI tools do this at massive scale and with tremendous sophistication. It’s also easy for our enemies to set traps by encouraging betting in areas they are trying to collect intelligence on.”
That’s why Rep. Eugene Vindman (D-VA), a veteran, sent a letter to Polymarket’s CEO, Sean Coplan, demanding the company’s preservation and disclosure of all records tied to bets on US military actions.
Framing the operational stakes in a television interview, Vindman said, “Whether it’s a couple hours or a full 24 hours, that’s an eternity in military operations — that’s plenty of time to either adjust troop dispositions or put air defense assets on guard.”
From predicting outcomes to shaping them
It doesn’t require widespread abuse to become a problem. A single insider generates signals others interpret. Sophisticated actors could manipulate markets directly — placing trades not to profit but to create the appearance of momentum, induce panic, or drive a policy reaction.
As Max Meizlish of the Foundation for Defense of Democracies put it: “If an adversary could make it seem that there’s instability in the market, that could actually have some really big downstream implications if the market more broadly starts to actually recognize that and operate based on those signals.”
Retired Gen. Mark Kimmitt views the Van Dyke case as something closer to treason than to insider trading. “What’s the difference between him monetizing this by walking in the nearest embassy and saying, ‘I’ve got classified material, you’ll want to buy it’ — and this?” he said to The Preamble. “This is just a new venue for doing it.” He added a sharper concern: “It turns these prediction markets from observers into influencers. Because if the people involved are trying to make money on it, they can influence the outcome.”
A prediction market trade is not the same as a classified briefing. An adversary cannot know with certainty whether a sudden shift reflects real knowledge, speculation, or deliberate misdirection. But in high-stakes environments, even the possibility of credible information can prompt defensive moves, miscalculations, or overreactions.
A regulatory gap
The Commodity Futures Trading Commission (CFTC) oversees these markets with roughly one-eighth the staff of the Securities and Exchange Commission (SEC), despite the fact that Kalshi alone receives more than $2 billion in trades every week. Van Dyke allegedly used cryptocurrency accounts to place his bets on Polymarket’s offshore platform. That version — which was set up outside of US regulatory reach and which, as part of a 2022 settlement, the CFTC ordered Polymarket to bar US users from accessing — has nonetheless been widely used to skirt those restrictions. That ban appears to have been lightly enforced at best.
The Trump administration has shown little appetite for tightening the rules. CFTC investigations into prediction markets have been dropped. Truth Social has announced plans for its own prediction market. Donald Trump Jr. serves as an adviser to both Polymarket and Kalshi, with his venture capital fund invested in Polymarket.
“Given the conflicted relationship of the first family,” said Yale’s Jeffrey Sonnenfeld, “CFTC oversight could be compromised.”
When Trump was asked about Van Dyke, he compared him to Pete Rose — who was banned from baseball for betting on his team’s wins. The comparison misses the point entirely. Van Dyke didn’t bet on a game everyone knew was being played. He bet on a classified operation no one outside a narrow circle was supposed to know existed — and, in doing so, left traces readable by anyone with the right tools.
The gamification of war
This is all unfolding alongside a broader change in the way we process news about conflicts. The Iran war has already demonstrated how fully both sides have embraced the information environment as a theater of operations. White House videos mix strike footage with scenes from the movie Top Gun and imagery from Call of Duty, the popular combat video game franchise. Iranian Lego animations on American social media feeds depict a cartoonish Trump figure fleeing missiles and watching helplessly as Iran prevails. Both sides understand that in 2026, the information space is no longer a backdrop to the conflict but a central part of it.
Prediction markets fit that same ecosystem. They convert the uncertainty of war into real-time metrics — a percentage chance of escalation, a contract on a leader’s survival. In this way, they become another layer of the gamification of war. Just as stylized video clips make destruction look like a highlight reel, betting markets make military outcomes seem like positions to be traded and opportunities to be exploited.
The open question
Van Dyke’s indictment has been greeted by some as proof the system works. Polymarket’s Coplan said the platform flagged Van Dyke’s account and cooperated with investigators. The problem is that by the time a trade is prosecuted, information has already moved through the system. Any adversary watching has already seen it.
The Defense Advanced Research Projects Agency (DARPA), the Pentagon’s research and development arm, tried something like this two decades ago — a government-backed exchange called the Policy Analysis Market, designed to price geopolitical contingencies and surface early intelligence signals. Congress shut it down in days, with one senator calling it a “federal betting parlor” on terrorism and assassination. What private industry has now built is far more advanced, far more liquid, and operating with almost none of the constraints that killed that earlier experiment. Every adversary in the world can access it.
If the allegations hold, Van Dyke made the mistake of being traceable. While Polymarket uses crypto wallets with pseudonyms, transactions are publicly recorded, and investigators can often link activity back to real identities through account data.
The next insider may be more careful. An adversary seeking to exploit these markets rather than profit from them may leave no trace at all. As Trump himself observed, “The whole world, unfortunately, has become somewhat of a casino.”
Without stronger efforts to regulate those in power from betting on the outcomes they control, the American people — and the national security apparatus sworn to protect them — could end up holding the losing hand.







