Prediction Markets Have an Insider Trading Problem: What Happens When Someone Already Knows the Future?

Prediction markets have spent the past few years trying to convince regulators, investors and the public that they are something more sophisticated than betting. The latest evidence from the United States makes that argument increasingly uncomfortable, because if platforms such as Polymarket and Kalshi really are markets, they may also have inherited one of the oldest problems in financial markets: insiders who know the answer before everyone else.

Newly disclosed records show that the U.S. Commodity Futures Trading Commission (CFTC) authorized at least three previously unreported investigations involving trading on Polymarket. The inquiries concern contracts linked to former President Joe Biden's pardons, Iran-related events and Google's 2025 Year in Search rankings. In the pardon markets, one trader reportedly made more than $300,000 after correctly anticipating several pardons. A network of accounts connected with Iran-related contracts reportedly generated around $2.4 million with an extraordinary success rate. The Google investigation comes alongside an already public CFTC enforcement action alleging that a Google employee used confidential information to generate approximately $1.2 million in profits from Year in Search contracts.

These cases have not all produced findings of wrongdoing, and an investigation is not proof of insider trading. But from an iGaming perspective, the pattern is now difficult to dismiss as a niche compliance issue. Suspicious prediction-market activity is appearing around politics, corporate information, technology, military and geopolitical events. At the same time, European regulators are openly warning about the same vulnerability. The question facing the sector is becoming much bigger than whether prediction markets constitute gambling or derivatives. It is whether a market built around predicting real-world events can remain fair when some participants may have access to the outcome before the rest of the market does.

The Fundamental Problem: What If the Outcome Is Not a Prediction?

There is a simple distinction between a casino game and a prediction market that deserves much more attention. In a properly regulated online casino, nobody is supposed to know the next roulette number, card or slot outcome. RNG-certified games are specifically designed to prevent that knowledge from existing. Operators, suppliers and testing laboratories spend enormous resources ensuring that outcomes cannot be predicted or manipulated.

Prediction markets invert that relationship. Their underlying events happen in the real world, where information is never distributed equally.

A political adviser could know that a pardon is being prepared. A corporate employee could see an unreleased product announcement. An auditor could potentially have access to earnings data before publication. A military or government employee might know that an operation is about to occur. A technology employee could know the result of an internal ranking before the company announces it.

For everyone else, the contract is a prediction. For the person holding the information, it may be closer to an answer sheet.

That creates what I believe is one of the most serious structural questions prediction markets will face in 2027: how do you price probability when some participants may already know the outcome?

Three Polymarket Investigations Change the Conversation

Documents obtained by WIRED through the Freedom of Information Act revealed that the CFTC authorized three private investigations into Polymarket-related trading during 2026. The first, authorized in May, concerned contracts involving pardons issued by Joe Biden. Reporting preceding the investigation had identified a trader who earned more than $300,000 after correctly predicting several pre-emptive pardons.

A second investigation concerned Iran-related event contracts. Reporting had previously identified a network of suspicious accounts that reportedly earned approximately $2.4 million from Iran-related positions with an extremely high success rate. The existence of an unusual trading pattern does not establish that confidential government or military information was used, but it illustrates exactly why geopolitical prediction markets create a surveillance challenge unlike anything encountered in conventional casino gaming.

The third inquiry relates to Google's Year in Search contracts and extends beyond an already significant enforcement case. In May 2026, the CFTC charged Google software engineer Michele Spagnuolo, alleging that he obtained sensitive non-public information concerning Google's official 2025 Year in Search results and traded at least 23 related Polymarket contracts with near-perfect accuracy. According to the CFTC, the activity generated approximately $1.2 million in profits. The U.S. Attorney's Office for the Southern District of New York simultaneously brought a criminal case based on similar alleged conduct.

CFTC Chairman Michael Selig made the regulator's position unusually clear when the Google action was announced:

“The Commission will not tolerate fraud, manipulation, or insider trading, regardless of the technology or platform that is used.”

That sentence may ultimately matter more to prediction markets than another argument about whether they resemble sportsbooks, exchanges or derivatives platforms. Regulators are increasingly looking beyond the label and towards what actually happens when money meets privileged information.

Then Came the KPMG Pattern

The latest corporate example makes the problem still harder to ignore. Blockchain-forensics company Bubblemaps identified 19 interconnected Polymarket accounts that had placed wagers around the earnings of companies audited by KPMG. According to reporting by The Wall Street Journal, the accounts won 41 of 42 relevant wagers, approximately a 98% success rate, across companies including Wells Fargo, Home Depot and DoorDash.

The network reportedly earned around $22,000 from those wagers. That is modest compared with some multimillion-dollar prediction-market stories, but the amount is not the important part. The pattern is.

Blockchain analysis reportedly linked the accounts through fund movements, raising the possibility that they were controlled by one person or a coordinated group. Bubblemaps stressed that its analysis was not proof of insider trading. KPMG has said it has zero tolerance for the misuse of confidential client information.

For market-integrity purposes, however, 41 successful positions out of 42 warrants attention regardless of the final explanation. Nicolas Vaiman, co-founder and CEO of Bubblemaps, described the pattern as,

“highly symptomatic of someone with an unfair information advantage.”

This is precisely where prediction markets begin to look less like casual wagering and more like markets requiring serious trade surveillance.

ESMA Has Now Entered the Debate

Europe is noticing the same problem.

The European Securities and Markets Authority [ESMA] has warned that prediction markets could possibly have insider trading potential, while highlighting the additional difficulties created when markets interact with crypto infrastructure and pseudonymous accounts. ESMA's concerns extend beyond insider information to manipulation, retail-investor protection and the gamification of financial-style trading.

This matters because prediction markets are simultaneously attempting to become more mainstream.

They increasingly appear alongside election coverage, economic analysis and financial-market commentary. Their implied probabilities are quoted as if they represent a collective assessment of the likelihood of an event. The attraction is obvious: unlike an opinion poll, participants are putting money behind their predictions.

But that perceived informational advantage becomes a weakness if the price is disproportionately influenced by someone who does not have a better prediction, but simply better access to confidential information.

A market can be extraordinarily accurate for the wrong reason.

Casinos Protect Players From Rigged Randomness. Prediction Markets Need to Protect Them From Non-Randomness

This is where the issue becomes particularly interesting.

Casino regulation has traditionally focused on making sure randomness really is random. Certified RNGs, game testing, audit trails, segregation of player funds, suspicious transaction monitoring and technical standards exist because a player needs confidence that the operator cannot secretly control the outcome.

Prediction markets have almost the opposite integrity problem.

They need to ensure that uncertainty really is uncertain.

If one participant already knows which executives will be pardoned, what Google's unpublished rankings contain, whether a company will beat earnings expectations or whether a government operation will begin tomorrow, the market has stopped functioning equally for all participants.

That does not necessarily mean the underlying contract should not exist. It means the surveillance architecture surrounding it has to become far more sophisticated.

Polymarket's Blockchain Is Both a Weakness and a Surveillance Tool

There is an interesting paradox here. Crypto-based prediction markets can make enforcement harder because users can operate pseudonymously, distribute positions across multiple wallets and move capital through complex blockchain structures.

But blockchain transparency can also expose behavior that would otherwise remain hidden.

The KPMG-related account cluster illustrates this perfectly. Investigators were able to analyze connections between wallets and trading behavior precisely because blockchain transactions leave persistent records. What looks like 19 independent traders at first glance can potentially be reconstructed as a connected network through funding flows.

So crypto does not simply make prediction-market surveillance impossible. It changes what surveillance should look like.

Platforms may increasingly need blockchain analytics combined with conventional financial-market surveillance: linked-wallet analysis, unusual win-rate detection, concentrated trading before announcements, IP and device intelligence, employee and politically exposed person screening where legally appropriate, abnormal account clustering and rapid escalation of trades linked to sensitive events.

That begins to resemble the surveillance infrastructure of a securities exchange much more than the risk department of an ordinary sportsbook.

The Kalshi-versus-Polymarket Regulatory Argument Is Becoming Secondary

Kalshi and Polymarket have spent considerable time at the center of arguments over whether event contracts belong under federal commodities regulation or state gambling law. That battle remains enormously important, particularly in the United States.

But insider information creates a problem whichever definition eventually wins.

If prediction contracts are financial products, regulators will reasonably ask why market-abuse standards should be materially weaker than those protecting traditional financial markets.

If they are gambling products, another question immediately follows: why should customers participate in a wager where another participant may legally possess decisive confidential information?

Neither answer is particularly comfortable.

The Google case captures this contradiction remarkably well. Lawyers for the accused employee have argued that the Polymarket transactions were effectively offshore bets rather than swaps regulated under the Commodity Exchange Act. That may become an important legal argument about jurisdiction, but from a consumer's perspective it leaves a much simpler question unanswered: should somebody be allowed to bet against me when they already know the result?

Prediction Markets Need an Insider-Trading Framework of Their Own

Simply importing Wall Street rules may not be enough.

Traditional insider-trading law developed largely around securities, issuers, fiduciary duties and material non-public information. Prediction markets can cover virtually anything: elections, wars, papal elections, celebrity events, economic statistics, technology launches, corporate results and government decisions.

Who counts as an insider in those markets?

For an earnings contract, the answer might include executives, accountants and auditors. For a political contract, it could include campaign staff or government officials. For a military event, it could potentially extend to soldiers, intelligence personnel, contractors or diplomats. For an entertainment market, it could include producers or employees involved in an unreleased result.

That means prediction-market compliance may eventually need event-specific restricted-person frameworks, rather than one universal definition of an insider.

Platforms could also impose stronger trading restrictions on markets where the outcome is determined internally by a small group of identifiable people. The probability of information asymmetry is fundamentally different between “Will inflation exceed 3%?” and “Which unreleased product will Company X announce tomorrow?”

The market design should recognize that difference.

The Bigger the Industry Gets, the Bigger This Problem Becomes

Prediction markets are moving rapidly towards mainstream finance, sports and political media. That growth makes insider-information controls more urgent, not less.

Small experimental markets can survive on novelty. Large markets involving millions of dollars need institutional credibility.

Once professional traders, retail customers, financial companies and major media organizations begin treating prediction-market prices as meaningful probability signals, market integrity becomes essential infrastructure.

If participants start believing that extraordinary trading results may come from confidential information rather than superior analysis, liquidity and trust can deteriorate quickly.

This is ultimately why the recent investigations matter.

The reputational danger is not simply that an insider could make $20,000, $300,000 or $2 million.

It is that ordinary users may begin to wonder who they are actually trading against.

What Happens When Someone Already Knows the Future?

Prediction markets have always marketed themselves around collective intelligence: thousands of participants combining information, judgment and financial conviction to estimate what will happen next.

That proposition is genuinely interesting. In many situations, market prices can aggregate information remarkably effectively.

But there is an enormous difference between forecasting the future and privately knowing it.

The newly disclosed CFTC investigations, the Google enforcement case, the unusual KPMG-related trading patterns and ESMA's increasingly explicit warnings suggest that insider trading is no longer a theoretical edge case for prediction markets. It is becoming one of the industry's defining regulatory challenges.

Polymarket, Kalshi and the wider sector therefore face a choice as they move towards 2027. They can continue debating whether they are betting companies, financial exchanges, information markets or something entirely new. Or they can recognize that their success has created obligations associated with all of those categories.

Because the industry's greatest competitive advantage, allowing people to trade almost any real-world outcome, is also its greatest vulnerability.
Casinos have spent decades building systems designed to reassure players that nobody knows the next result.
​​​​​​​
Prediction markets may now have to build systems capable of detecting the people who do.

References

Commodity Futures Trading Commission. (2026, May 27). CFTC charges Google employee with insider trading in search result-related event contracts. U.S. Commodity Futures Trading Commission. CFTC enforcement release
Knibbs, K. (2026, September 11). The US government launched 3 previously unreported investigations of Polymarket trades. WIRED. WIRED investigation
Reuters. (2026, September 10). EU watchdog flags risk of abrupt market correction. Reuters. Reuters report
The Wall Street Journal. (2026, September 12). Cluster of Polymarket accounts won big on companies audited by KPMG. The Wall Street Journal. Wall Street Journal investigation
Financial Times. (2026, September 10). Prediction markets “rife with insider trading”, warns EU watchdog. Financial Times. Financial Times report on ESMA warning
Financial Times. (2026, August 31). It's not insider trading, it's offshore gambling! Financial Times. Financial Times analysis of the Google case