What counts as "inside information" in a prediction market?
That question sits at the heart of a significant enforcement action brought by the Commodity Futures Trading Commission (CFTC) against a Google employee accused of using confidential company data to generate more than USD 1 million in profits on Polymarket.
On the surface, this might look like a relatively contained enforcement action against one individual. In context, it is anything but. The case may become one of the clearest tests yet of whether prediction markets are simply sophisticated betting platforms or federally regulated derivatives markets where insider trading rules apply with full force.
More fundamentally, it raises a question that compliance professionals, regulators and market participants are only beginning to confront: when markets allow participants to trade on virtually any event with economic consequences, how far does the concept of "inside information" extend?
What Happened
On 27 May 2026, the CFTC filed a civil complaint in the Southern District of New York against Michele Spagnuolo, a Swiss-resident software engineer employed by Google. On the same day, the U.S. Attorney's Office for the Southern District of New York unsealed a parallel criminal complaint. The charges: insider trading on the prediction market platform Polymarket using non-public information about Google's annual "Year in Search" list for 2025.
According to the CFTC's complaint (CFTC v. Michele Spagnuolo, No. 26-cv-4419, S.D.N.Y.), Spagnuolo (trading under the Polymarket handle "AlphaRaccoon") purchased "Yes" or "No" shares in at least 23 event contracts linked to Google's Year in Search rankings between approximately October and December 2025. The contracts covered questions such as whether Pope Leo XIV, Kendrick Lamar or Zohran Mamdani would rank among the top-searched individuals of the year; whether Squid Game would be the most-searched TV show; and whether Sydney Sweeney would be the most-searched actor, among others.
Spagnuolo correctly predicted 22 of his 23 positions, generating approximately USD 1.2 million in profits. His Polymarket account received roughly 3.9 million USDC.e upon contract resolution, with proceeds subsequently routed through a cryptocurrency swapping service to an account in his name, opened with his Italian government identification.
His access to the underlying data was straightforward. As a Google employee, Spagnuolo had allegedly accessed confidential internal tools showing the anticipated Year in Search rankings on at least two occasions (15 October and 27 November 2025) well before Google's public release of the list on 4 December 2025. Google's corporate policies required employees to maintain the confidentiality of such proprietary information, and Spagnuolo had completed training on those policies.
The complaint charges violations of Section 6(c)(1) of the Commodity Exchange Act and CFTC Regulation 180.1(a)(1) and (3) — the same anti-fraud and anti-manipulation provisions that underpin the CFTC's broader insider trading framework. It seeks disgorgement, restitution, civil monetary penalties and permanent trading and registration bans.
Why "Year in Search" Data Qualifies as Material Non-Public Information
This is where the case takes on broader significance.
The sensitive information at issue is not what most compliance professionals would instinctively recognise as "inside information." There are no earnings figures, no M&A discussions, no client order flow and no government policy decisions. It is, on its face, a list of popular search terms.
The obvious question is: why does this matter commercially?
The complaint implicitly reflects the view that the information carried commercial value beyond mere curiosity. Google generates the overwhelming majority of its revenue from advertising products leveraged through its search platform. The Year in Search campaign is a deliberately staged annual marketing event designed to generate media coverage and demonstrate the scale of Google's search platform to existing and prospective advertisers. Foreknowledge of the rankings therefore carries genuine commercial value: it predicts, in effect, what Google's marketing machine will amplify, and to whom.
That framing is important. The CFTC is not merely asserting that the information was confidential. It is asserting that the information had financial, economic or commercial consequences (the statutory threshold under the Act for event contracts to constitute swaps and therefore fall within federal jurisdiction).
The Jurisdictional Stakes: Swaps, Not Gambling
This is the central regulatory argument the CFTC has been building across multiple fronts, and the Spagnuolo case sharpens it considerably.
Under Section 1a(47) of the Commodity Exchange Act, a "swap" includes any agreement, contract or transaction providing for a payment dependent on the occurrence or non-occurrence of an event or contingency associated with a potential financial, economic or commercial consequence. The CFTC's position (argued in the complaint and in a growing body of litigation) is that event contracts traded on platforms like Polymarket meet this definition, bringing them within the CFTC's asserted federal jurisdiction and outside the gambling and gaming frameworks that individual U.S. states have traditionally administered.
The implications of that jurisdictional claim are significant. If event contracts are swaps, then trading in them while in possession of material non-public information is not merely a terms-of-service violation or a matter of internal corporate policy. It is federal market manipulation and fraud, subject to the full weight of the Commodity Exchange Act and CFTC enforcement authority.
The CFTC has been pressing this theory aggressively. To date, the agency has filed cases directly against at least seven states (Arizona, Connecticut, Illinois, Minnesota, New York, Rhode Island and Wisconsin). The cases have sought to restrict or prohibit prediction market activity under state gambling laws. It has also joined actions in Maryland, Nevada and Ohio, and has filed amicus briefs in at least two appeals.
The battleground is whether these markets are federal derivatives markets or state-regulated gambling. Given the growing number of federal and state challenges, a circuit split remains a realistic possibility, potentially setting the stage for eventual Supreme Court review.
The Spagnuolo case adds a different dimension. It is not about market classification per se, but about the conduct of individual participants once the market-as-swap framing is accepted. In that sense, it serves as a proof of concept for the CFTC's enforcement theory.
The Expanding Universe of "Inside Information"
This is where the implications extend well beyond Google.
The Spagnuolo matter is emblematic of a structural challenge that prediction markets pose to traditional insider trading frameworks. Conventional securities law insider trading concerned a relatively bounded category of information: corporate earnings, undisclosed M&A, a client's trading intentions or, in some cases, government-held policy information. Case law developed over decades to address a relatively stable informational ecosystem.
Prediction markets dissolve that stability.
When the subject matter of a contract can be anything with potential commercial consequence, the universe of potentially material non-public information expands dramatically.
Consider the cases already in or near the enforcement pipeline: trading on knowledge that a particular athlete is injured before that information is public; positioning on geopolitical event contracts based on access to classified government assessments; or, as here, trading on corporate marketing data that would be unremarkable in almost any other context.
The question increasingly becomes not whether information is traditionally viewed as market-moving, but whether it is capable of determining the outcome of a tradable event contract.
The Spagnuolo case is not the first to explore this territory. The CFTC has previously brought and settled actions against Polymarket itself. In January 2022, Blockratize Inc. agreed to pay USD 1.4 million and wind down unlawful operations in the United States for operating an unregistered swap facility and offering off-exchange binary options contracts.
More recently, concerns have been raised (though not always resolved through formal enforcement) around whether prediction market positions in political event contracts could constitute a form of market manipulation when taken by actors with privileged access to polling, campaign, opposition research data or even if the actor themselves will or will not do something.
The common thread is misappropriation: the use of information obtained under a duty of confidence for personal gain in a market where that information is price-determinative.
What This Means for Compliance and Surveillance Professionals
For the bank or broker surveillance officer, much of the current regulatory landscape may still feel academic.
Prediction markets are largely accessible without intermediaries. Participants transact directly through DeFi protocols, wallets and on-chain mechanisms that sit outside traditional brokerage infrastructure. There is, as yet, no readily apparent hook for existing trade surveillance frameworks.
That position is unlikely to remain tenable for long.
Regulated broker-dealers and exchanges are already exploring whether and how to offer access to event contracts through conventional account structures, which would bring them within existing conduct-of-business, AML and market abuse frameworks. CFTC Chairman Michael Selig has made clear that the Commission views prediction markets as within its remit and intends to treat them accordingly.
When that infrastructure arrives, the surveillance challenges will be substantial.
Detecting insider trading in prediction markets requires monitoring not just the trade itself, but the correlation between a position and the availability of non-public information (across a far broader range of potential information categories than traditional securities surveillance contemplates).
An employee of a sports analytics firm, a government contractor or a technology company with proprietary data sets could all, in principle, possess information material to an event contract. The "inside information" in such cases will not announce itself with a press release or a filing date.
That is precisely why the Spagnuolo case matters.
Its significance may ultimately have little to do with Google itself.
Instead, the case offers an early indication of how regulators may approach insider trading in a world where markets increasingly allow participants to trade on almost any event with economic consequences. The traditional boundaries of material non-public information were developed around securities markets and a relatively predictable set of information sources. Prediction markets challenge those assumptions.
For compliance, surveillance and risk professionals, the implications extend far beyond any single platform or enforcement action. If prediction markets continue to expand, firms and regulators will need to develop new ways of identifying, monitoring and governing informational advantages that may originate far outside the categories traditionally associated with insider trading.
The industry will also need to continue grappling with a fundamental threshold question: are these markets best understood as gambling arrangements, or as financial instruments subject to the full framework of federal market regulation? The answer will shape not only market structure and participant obligations, but also the future scope of market abuse enforcement.
The CFTC has made its position increasingly clear. The courts now have the opportunity to determine how far that position extends, and, in doing so, help define what "inside information" means in the age of prediction markets.
CFTC v. Michele Spagnuolo, No. 26-cv-4419 (S.D.N.Y., filed 27 May 2026). CFTC Press Release No. 9237-26.