A former U.S. congressman who posted a video promising to attend President Trump's State of the Union address, then quietly bet against his own attendance and skipped the event, is now under federal investigation for market manipulation.
Kalshi, the regulated prediction market where the trades occurred, detected the activity, froze George Santos' account, and referred the case to both the Commodity Futures Trading Commission and the Department of Justice. Both agencies opened investigations into Santos, according to two people familiar with Kalshi's internal probe who were not authorized to speak publicly, NPR reported on June 2, 2026.
Santos' alleged scheme ran on a simple logic: create a public signal, let other traders move the market on it, then collect the difference when reality diverges from the lie.
The play. On February 23, 2026, Santos posted a video to X confirming he would be in the gallery for Trump's address. "I'm going to be there for the State of Union in the gallery, guys," he said. Traders on Kalshi were placing millions of dollars in bets on who would attend — Santos' video sent the odds for his attendance soaring. He had already placed bets predicting he would not appear, according to three people with direct knowledge of his trades who were not authorized to speak publicly. As Trump spoke, Santos was at an airport. "Watching SOTU from an airport tv was not part of the plan! FML," he posted to X. The odds collapsed. Santos walked away with a profit in the tens of thousands of dollars.
Santos' response. Reached by NPR, Santos said "Well, that's news to me" when asked about the investigation. When asked whether he had a Kalshi account, he said, "I'm not saying yes, I'm not saying no." He told NPR that Kalshi co-founder Luana Lopes Lara is "a fellow Brazilian" he knows personally and said he would call her about the investigation. He did not follow up. A person familiar with Kalshi's investigation told NPR that Santos does not know Lara. Kalshi has asked to interview Santos as part of its probe; he has dodged those requests.
Why this case is legally novel. Classic insider trading involves exploiting confidential information — an earnings announcement, a merger, a nonpublic government decision. Santos had no confidential information. He was the information. Yuriy Brisov of Digital & Analogue Partners told Decrypt the case is "closer to manipulation" than insider trading, and that "trading on your own conduct is a category that the inherited rulebook never anticipated." The CFTC's anti-manipulation authority under the Commodity Exchange Act covers false or misleading statements designed to affect a market price. Whether a public tweet qualifies as the kind of intentional deception that meets that threshold is the core legal question regulators will have to answer.
Santos' record. Santos served four months of a seven-year federal sentence before Trump commuted it last October. He had been convicted of wire fraud, money laundering, and stealing from political donors after a campaign built on fabrications about his academic record, his family's Holocaust history, and his professional background. "Virtually everything that he put out about himself when he was running for office was manufactured," Case Western Reserve law professor Jonathan Entin told NPR. The pattern prosecutors will argue they are seeing now is not new.
Widening federal scrutiny of prediction markets. Santos is the highest-profile name in a string of prediction market enforcement actions that accelerated through spring 2026. In April, federal prosecutors criminally charged a U.S. Army Special Forces soldier accused of making more than $400,000 betting on the capture of Venezuelan leader Nicolás Maduro on Polymarket. Last week, the DOJ charged a Google engineer for making more than $1 million betting on search trends using confidential information from his employer, also on Polymarket. The House Oversight Committee opened an investigation into both Kalshi and Polymarket in May 2026.
What has drawn attention in each case is how quickly the platforms themselves detected the activity — before regulators did. Kalshi identified Santos' trades, froze his account, and made referrals on its own initiative. The platforms' surveillance capacity reflects a structural incentive absent from traditional markets: if manipulation becomes routine, bettors stop trusting the odds, and the market dies. That self-interest in market integrity has made Kalshi and Polymarket more aggressive investigators than the agencies nominally overseeing them.
What it means for regulation. Congress is debating whether to give prediction markets a permanent legal framework in the United States. The Santos case will inform that debate in a way that abstract arguments about market structure cannot. The question is no longer whether manipulation can happen in these markets — it is whether the law, written before public figures could tweet themselves into a position and then profit from the crash, is adequate to address it. The Santos investigation suggests regulators are treating it as a live question.
Primary source: NPR investigation published June 2, 2026. Santos' X posts published February 23 and February 24–25, 2026 (during the State of the Union address). Kalshi referral and DOJ/CFTC investigation opening per NPR's reporting on anonymous sources with direct knowledge. Legal analysis per Yuriy Brisov (Digital & Analogue Partners) as quoted by Decrypt.