A Google software engineer has been accused of leveraging insider information to net more than $1 million on the decentralized prediction platform Polymarket. According to an unsealed criminal complaint in New York, 36-year-old Michele Spagnuolo, an Italian citizen residing in Switzerland, "misappropriated confidential and valuable nonpublic information from his employer and used that information to place a series of Google-related bets on Polymarket."
The complaint alleges that between October 2025 and December 2025, Spagnuolo placed highly targeted wagers tied to Google’s most-searched people of 2025. Specifically, he bet that high-profile figures such as Donald Trump, Pope Leo XIV, and Kanye West’s wife Bianca Censori would not take the top spot. Conversely, he bet that the recording artist D4vd would secure a top-five position and eventually finish first. These informed bets ultimately generated approximately $1.2 million in profit.
Federal prosecutors emphasized that Spagnuolo held an unfair advantage over everyday traders, knowing the outcomes beforehand by accessing Google's confidential internal marketing data. Spagnuolo now faces charges of commodities fraud, wire fraud, and money laundering, which carry potential prison sentences ranging from 10 to 20 years. Concurrently, the Commodity Futures Trading Commission (CFTC) has filed a complaint accusing him of insider trading.
A Google spokesperson stated that the employee accessed the internal marketing tool—which was technically accessible to all employees—but using confidential data to place wagers is a severe breach of policy, adding that the employee has been placed on leave. Polymarket stated it worked closely with the DOJ and CFTC, boasting that it is the first prediction platform whose cooperation directly led to US insider trading charges. This case follows the recent arrest of a US soldier who won $400,000 betting on the removal of Venezuelan President Nicolás Maduro, highlighting mounting scrutiny on prediction markets like Polymarket and Kalshi.
[AgentUpdate Depth Analysis] This case highlights a critical vulnerability at the intersection of prediction markets and the burgeoning AI Agent ecosystem. As autonomous AI Agents are increasingly deployed to execute real-time, high-stakes trades on decentralized protocols like Polymarket, the definition of "insider trading" will inevitably expand. If AI Agents leverage advanced web scraping, semantic analysis of private APIs, or proprietary databases to outmaneuver human retail traders, the systemic trust in prediction markets could collapse. This incident underscores the urgent necessity for robust data governance and "AI Guardrails" within enterprise systems. Developers must implement strict data access boundaries, as agentic workflows capable of exploiting asymmetric data will rapidly turn into autonomous, untraceable insider-trading engines. The future of decentralized prediction relies heavily on how effectively the tech industry can police automated access to proprietary real-time data.
