Author: Polyfactual
Compiled by: Hu Tao, ChainCatcher
In late February 2026, four anonymous wallets appeared on the Polymarket platform. These wallets had been created only days before and seemed brimming with confidence. Over the next few weeks, they placed over 80 bets on specific mechanisms of a US-Iran war, the timing of the first strike, the removal of Iran's Supreme Leader, and the announcement of a ceasefire. When Bubblemaps finally mapped this cluster of bets and linked the initial four wallets to five additional ones, it was discovered that these nine associated accounts had collectively won over $2.4 million in winnings with a win rate of 98%, despite many of the bets being placed at long odds.
This phenomenon now has a name, or at least a category: Information Laundering. To understand why it's so corrosive, one must first understand the nature of prediction market prices, because the mechanisms that make these markets function are the same ones that make them vulnerable to exploitation.
Stripped of its crypto wrapping, a PM contract is quite simple. Each share pays out $1 if correct and nothing if wrong. Because each binary question has only two outcomes, one YES share plus one NO share always equals $1. Therefore, a YES share priced at $0.36 indicates the market believes the prediction has a 36% chance of being correct.
Crucially, Polymarket doesn't set these prices. They originate from a trader order book (CLOB). Supply and demand among traders determine the price, and the displayed price is at the midpoint of the bid-ask spread. Herein lies perhaps its genius. In this model, the price isn't a bookmaker's opinion; it's the collective expectation of all traders in the order book. When new information emerges, like a strong jobs report or lower-than-expected CPI data, traders reprioritize, and the price adjusts accordingly. Effectively, the market becomes a continuously updating probability estimate, and financial institutions are willing to pay for this. Institutions like Bloomberg, Reuters, and hedge funds now purchase real-time access to Polymarket data feeds, viewing them as a faster market sentiment indicator than traditional polls.
The trap, however, is that a system designed to transform information into price cannot distinguish between public information and stolen information. The order book doesn't ask where your edge comes from; it just registers that you bought.
This is where the term "laundering" becomes apt. In traditional money laundering, dirty cash flows in one end, and clean, untraceable cash flows out the other. In information laundering, confidential information flows in one end, and a market price—bearing no trace—flows out the other.
For example, suppose someone knows a strike will happen in 48 hours, while the market currently prices it at 15%. Their buying pressure consumes all the sell orders in the order book, pushing the midpoint price up, say, to $0.35. To everyone else, this looks like a normal repricing, as if a trader made a sharp geopolitical call. The secret is neatly packaged into a clear signal. When the strike happens, YES contracts pay $1. Positions bought around $0.15 yield roughly a 6.7x return. The Maduro case from months ago illustrated this scale clearly. Prosecutors alleged the Army Sergeant turned roughly $34,000 in bets into about $400,000.
The laundering metaphor also fits the cover-up. Bubblemaps found the Iranian crime ring's losses were tiny, only a few hundred dollars, which the company believes were intentional to mislead investigators. A 98% win rate looks superhuman, but a 98% win rate with some trivial, deliberate losses looks almost like a very good trader.
Yet, the supreme irony is that these markets are more transparent than traditional exchanges. Even if account holders stay anonymous, every transaction is at least recorded on a public ledger. It's this openness that allows analysts using tools like Bubblemaps to reconstruct a nine-wallet conspiracy based on temporal correlation and volume, such as trades logged days before a market move on Feb 28th.
But this same transparency creates a secondary risk that deeply worries regulators. If outside analysts can decipher a coordinated group heavily betting on an attack, so can adversaries. Adversarial observers could spot anomalous trades and adjust their war plans and market predictions accordingly. An unusual spike in a certain war market is a low-cost, deniable intelligence source for anyone watching the chain. The launderers clean their information, and as a byproduct, broadcast the original secret to the world in abstracted form.
Why doesn't existing law simply cover this? Because traditional insider trading rules are built around stocks, material non-public information related to companies, earnings, M&A, executive disclosures, etc., not the timing of military operations. War has no "issuer," and no corporate insiders in the legal sense.
The geography of jurisdictions compounds the problem. U.S. federal law prohibits prediction markets from offering bets on wars or assassinations, but Maduro's bets were placed on Polymarket's offshore site, free from these constraints. And the barrier to entry is laughably low, easily circumvented by a $2-a-month VPN to bypass the U.S. ban. A KYC'd account is also simply for sale. Nonetheless, Washington is finally paying attention. On May 22nd, the House Oversight Committee launched a formal inquiry into prediction markets, demanding records on how they verify identities, enforce geographic restrictions, and handle suspicious trades related to Venezuela and Iran. Proposed bills, the 'No Death Bets Act' and the 'Financial Prediction Market Public Integrity Act,' aim to ban war betting and prohibit officials from trading using non-public information.
The brutal reality is that information laundering isn't a manufactured glitch in prediction markets; it's a side effect of their core operating mechanism. A market that perfectly transforms knowledge into price is, by its nature, going to reward those with the best information, including those who shouldn't have it. The hole can't be fully plugged without crippling the very mechanism that makes these markets more accurate than polls.
As the industry looks ahead, even adoption by just 1-2% of derivatives traders could push annual volume to $50 billion. The question is no longer if prediction markets work, but that they work too well. The question is whether a society can tolerate a machine that transforms its most closely guarded secrets into public, quotable, tradable numbers—and handsomely pays those who hold them.
