FBI Claims Google Employee Won Over $1M On Polymarket Bets Using Company Data

In a case of new-age insider trading, a Google software engineer has been accused of using company data to wager millions on the prediction market Polymarket, where visitors can gamble on everything from Paul Thomas Anderson's Oscar win to international conflicts.  Michele Spagnuolo, an Italian software engineer residing in Switzerland, is accused of using his position within the company to wager $2,754,092 on Google's "Year in Search" data. Accused of wire fraud, money laundering, and commodities fraud, Spagnuolo's charges could carrying up to 50 years in prison. Spagnuolo was arrested and charged before a New York federal court last week. A civil complaint from the Commodity Futures Trading Commission has also been lodged against the Google engineer.

In a statement announcing the criminal complaint, U.S. Attorney Jay Clayton said, "Today's charges reinforce a decades-old message: corporate insiders cannot use confidential business information to turn a profit in our markets." Clayton alleged that Spagnuolo "used Google's confidential business information to make more than $1.2 million in trading profits on Polymarket. Insider trading compromises the integrity of our markets, and the American people want this greed-driven conduct investigated and prosecuted." 

Prediction markets have seen a slew of high-profile cases in recent months. In April 2026, the DOJ charged Army Master Sgt. Gannon Ken Van Dyke with using inside information to wager on the abduction of Venezuelan President Nicolas Maduro. Other suspected instances have involved congressional elections, presidential pardons, and presidential announcements. One disgraced congressman is under investigation for wagering on his own appearances. Some experts have labeled the phenomenon an existential crisis for the booming industry. Others argue the trend is essential to the market's business models. As prediction sites are incorporated into everything from news broadcasts to financial applications like Google Finance, such concerns are paramount.

According to the court's criminal complaint against Spagnuolo, the 36-yearold software engineer used Google's internal software tool to access its 2025 Year in Search data in October and November 2025. Using this confidential information, Spagnuolo allegedly wagered $2,754,092 across 25 Google Year in Search outcomes under his Polymarket account, "AlphaRaccoon." The scheme allegedly netted an improbable $1.2 million in profit. 

Spagnuolo's downfall began with an X post, and is endemic of both the difficulty of policing prediction markets and the collective efforts to police them. Because Polymarket accounts do not require identification, Spagnuolo's activity was wholly anonymous. However, similar to cryptocurrency platforms, all trades are publicly logged via their blockchain. This dichotomy, in which trades are publicly visible white the identities of traders remain private, has sprung a cottage industry of prediction market internet sleuthing. One such online sleuth, a blockchain engineer and prediction market enthusiast, Haeju Jeong, was the first to notice Spagnuolo's suspicious trading history. Noting that "AlphaRacoon" went 22-for-23 on his predictions, Jeong concluded in an X post that "AlphaRacoon" was "a Google insider milking Polymarket for quick money." According to the complaint, Spagnuolo's Polymarket username was "promptly removed" after speculation began to circulate online.

In withdrawing his winnings from Polymarket, Spagnuolo further obfuscated his identity through a series of cryptocurrency swapping and transfer services. As with prediction market wagers, crypto transactions are typically logged via public wallet identification numbers, which allowed the FBI to map Spagnuolo's cryptocurrency money trail. Ultimately, the FBI discovered Spagnuolo's identity by following this digital record to the payment processor Nexo, where the Google employee opened an account using his Italian ID card. 

Insider trading may threaten prediction markets' survival. According to Kalshi, the largest American prediction market, insider trading investigations surpassed their 2025 totals by the first quarter of 2026(via ABC). Instances ranged from the results of Survivor to Israeli missile strikes. In May 2026, data analytics firm Bubblemaps found that nine interlinked accounts netted $2.4 million through 80 bets on U.S. military actions in Iran with 98% accuracy (via CBS). Political and military wagers are of particular concern. Although Polymarket is banned in the U.S., users easily circumvent rules through VPNs and cryptocurrency. Reportedly, more than $1 billion have been wagered on military actions in 2026, including on the fate of the U.S. airman down in Iran in April.

Policing prediction markets is difficult. Subject to less regulation than either Wall Street or online gambling platforms, prediction markets are regulated by the Commodity Futures Trading Commission (CFTC). The Trump administration has taken a hands-off approach, shrinking the CFTC's staff by 25%, filling only one of the commission's five seats, and suing states that attempt to regulate the industry. And while pushing for regulation has been a bipartisan endeavor, so too has the effort to hamstring enforcement.

Prediction market executives claim to prioritize cracking down on insider trading. However, its fair to question the industry's commitment. In interviews, Polymarket's founder Shayne Coplan has called insider trading "an inevitability" that "creates this financial incentive for people to go and divulge the information to the market." This speaks to Coplan's position that prediction markets are "truth machines," where anonymized, free market principles run free. However, these statements underscore the centrality of insider trading to the industry's perceived value. The bigger question may not be whether prediction markets can stop insider trading, but whether they aim to.