Prediction Markets Under Fire: Criminal Prosecutions, Moral Hazard, and the Exploitation of a $25-Billion Sector

By Adrian Cheek, Senior Cybercrime Researcher

In January, a US Army master sergeant allegedly helped plan the operation to capture Venezuelan leader Nicolas Maduro. Then, prosecutors say, he bet $33,000 on it happening and walked away with more than $400,000. Two months later, a Google engineer was charged with using internal search-trend data to win $1.2 million on prediction markets. These are the first federal criminal prosecutions ever brought against prediction market traders, and they landed within five weeks of each other. 

Prediction markets went from an academic curiosity to a multi-billion-dollar slice of the financial system in under two years, and the threat picture has moved just as fast. The most important change since the start of the year is simple to state. Meaningful criminal enforcement activity has emerged. For most of the sector’s life the only penalties for insider trading on these platforms were token fines the platforms handed out to themselves. As of late May there are two federal criminal prosecutions in the courts, the first of their kind, and one of them is against a serving Army master sergeant mentioned above. Those cases sit in the foreground of this assessment, because they change the deterrence calculation more than anything else on the table.

Running alongside the criminal cases is a jurisdictional fight between Washington and the states that has become genuinely messy. The CFTC is suing six states. One federal appeals court has sided with the platforms while another leans the other way. The first outright state ban was signed into law in May and sued over inside 24 hours. And the agency’s first comprehensive federal rule is now sitting on the President’s desk for review. None of it is settled. If anything the legal status of large parts of the sector is more uncertain now than it was six months ago.

Everything below assumes a reader who already knows these markets exist. The mechanics are covered quickly; the weight is on what can go wrong and what is now being done about it. Where a statement is an analytic judgment and not a matter of record, the text says so and gives its confidence level.

Key Findings About Prediction Markets

• Insider trading on prediction markets is no longer theoretical. Two federal criminal prosecutions (the first of their kind) were unsealed in spring 2026, one involving a military insider betting on an operation he helped plan, the other a corporate employee exploiting proprietary data. Both carry felony charges including wire fraud, commodities fraud, and money laundering.
• The regulatory landscape is more fractured than at any point in the sector’s history. The CFTC is suing six states, two federal appellate courts have reached opposing conclusions on preemption, the first state ban was signed into law in May 2026, and the agency’s first comprehensive federal rule is sitting on the President’s desk. Nothing is settled.
• Markets on war, assassination, and political leadership create genuine moral hazard. More than $1 billion was wagered on the US-Israeli campaign against Iran in early 2026. Contracts on military operations and the survival of heads of state fuse financial speculation with operational risk in ways no existing regulatory framework was designed to handle.
• AI-enabled actors now dominate price formation. By mid-March 2026, 14 of Polymarket’s 20 most profitable accounts were bots. Autonomous trading agents exploit latency, run wash trades across wallet clusters, and move faster than any current surveillance system can flag them.
• The global regulatory consensus is moving against the sector. By mid-2026, Polymarket faced restrictions in more than 33 jurisdictions. India, Spain, Indonesia, and multiple EU member states have independently concluded that prediction markets are gambling, regardless of the underlying technology.

The Instrument, Briefly

An event contract is a derivative with a binary payoff: it pays $1 if a specified event happens and $0 if it does not, with the price in between read as the market’s implied probability. A contract at $0.72 implies a 72% chance. Unlike ordinary derivatives, these are anchored not to a stock or commodity but to a discrete, observable outcome, an election result, a sports score, an interest-rate decision, and increasingly a military action or the survival of a head of state. That last category is the source of most of the problems in this report.

Two platforms dominate:

• Kalshi is a CFTC-registered designated contract market running a centralized order book inside the United States. 
• Polymarket runs a split operation: a regulated US exchange and a much larger offshore exchange on the Polygon blockchain, the latter based in Panama and nominally closed to American users, though VPNs make that a soft barrier. 

Both are well funded, the operator of the New York Stock Exchange committed up to $2 billion to Polymarket in October 2025, Kalshi raised $1 billion that December, and a crowd of entrants has followed: Robinhood, Coinbase, Crypto.com, FanDuel, DraftKings, PrizePicks, Underdog, and Fanatics among them. Sector volume passed roughly $25 billion in a single month in early 2026 and has kept climbing.

One figure cuts against the “wisdom of crowds” pitch. A Wall Street Journal review of 1.6 million Polymarket accounts found that 0.1% of them, fewer than 2,000 traders, took 67% of all profits while more than 70% of users lost money. The winners are overwhelmingly professional firms running algorithms against premium data feeds. This raises questions regarding the extent to which these prices reflect broad participant sentiment and how far they reflect a few firms’ edge, which bears on both the integrity argument and the consumer-protection case the states keep making.

Insider Trading is No Longer a Theoretical Problem

The structural issue is easy to state. These markets create a direct way to cash in non-public information, and the legal apparatus for stopping it is a fraction of what exists for securities. Insider-trading law under SEC Rule 10b-5 requires the purchase or sale of a security, and most event contracts are not securities; they are commodity derivatives, policed (if at all) by the CFTC under the Commodity Exchange Act. Rule 180.1 offers a parallel antifraud provision, but until this spring it had never been tested against a prediction market, and prosecutors have had to improvise with wire fraud, money laundering, and state commodities-fraud statutes like California’s Commodity Law and New York’s Martin Act.

Then two cases landed in quick succession, both out of the Southern District of New York, the office that had spent the early part of the year signaling that the “prediction market” label bought no immunity.

The Van Dyke Case (Unsealed April 23)

Army Master Sergeant Gannon Ken Van Dyke was charged with five felonies. Prosecutors allege he helped plan the operation to capture Venezuelan leader Nicolas Maduro, then used what he knew to put roughly $33,000 into Polymarket contracts on the raid and walked away with more than $400,000 when Maduro was taken on January 3. 

These allegations are assessed as particularly significant from a national-security perspective because they represent a direct intersection between operational access and financial incentive, an insider allegedly betting on an operation in which he was involved.

The Spagnuolo Case (Announced May 27)

Michele Spagnuolo, a Google security engineer and Italian national living in Switzerland, was charged with commodities fraud, wire fraud, money laundering, and related counts. The allegation is that he used internal Google data on search trends to make about $1.2 million betting on which individuals would top Google’s most-searched list for 2025, with the CFTC piling on a civil suit for disgorgement and penalties. 

These allegations may have broader implications for corporate compliance and insider-risk management, because it pulls the corporate world in. Contracts now track public-company events directly, stock moves, earnings-call language, regulatory rulings, product launches, and an employee trading on internal knowledge of any of those is doing something functionally identical to securities insider trading. Most corporate compliance manuals say nothing about prediction markets. If that case holds up, that is going to have to change.

Assessment

Both sets of charges are allegations not yet tested at trial, and the analysis below treats them as such.

This is the firmest judgment in the assessment, and confidence in it is high. Insider trading is occurring on these platforms, and it is no longer seriously arguable. The case rests on documented platform enforcement actions, on-chain trading patterns, and two federal indictments. Everything else here is built on top of it.

These indictments do not settle very much in themselves, but what they do settle matters. Ordinary fraud statutes can reach event-contract trading in practice, not just on paper. Whether they produce convictions is the open question. A guilty verdict or a meaningful plea in either case would almost certainly open the door to charges around Iran and election activity catalogued below.

The Pattern Around the Prosecutions

The prosecutions followed a series of incidents that had already attracted regulatory, legislative, and public attention. They sit on top of a year of incidents that drew the attention of regulators, lawmakers and, eventually, the public:

• Candidates betting on themselves: A candidate caught trading his own race on Kalshi in May 2025 drew a $2,246.36 penalty and a five-year ban. By April 2026, Kalshi had fined and suspended several more federal and state candidates for the same thing, in California, Minnesota, Texas and Virginia among others, and it now bars members of Congress from holding accounts at all.
• The content insider: A video editor for a major YouTube operation traded contracts on what his own channel would publish next, with advance knowledge of the content. Kalshi’s surveillance flagged a near-perfect record in low-odds markets; the penalty was $20,397.58 and a two-year ban, still the largest platform fine on record and a figure that tells you how little deterrence the platforms can muster on their own.
• Iran: Polymarket’s offshore book saw a wave of large bets on U.S. strikes just before they happened in February. One account under the handle “Magamyman” cleared $553,000, including a $32,000 wager placed hours before the strikes when the market gave them only a 17% chance. A Bubblemaps analysis aired by 60 Minutes traced nine accounts that together made $2.4 million timing key moments of the conflict, and individuals have been arrested in Israel for allegedly trading the same contracts on classified knowledge of strike timing.
• The Khamenei contract: A Kalshi market on whether Iran’s Supreme Leader would be removed drew $54 million in volume before it was paused. It was resolved when Khamenei was killed in the U.S.-Israeli strikes on February 28, and the resolution caused a fight, because Kalshi’s rules treated death as distinct from the “ouster” traders thought they had bet on and settled on the last traded price before death was confirmed.

Platform and Congressional Response

Both platforms have scrambled to get in front of this. Polymarket rewrote its rules in March to bar trading on breaches of trust or by anyone in a position to influence an outcome; Kalshi has tightened the screws on officials and candidates. Whether self-policing means much when the largest fine anyone has paid is roughly $20,000 against potential profits in the hundreds of thousands is exactly the question Congress is now asking.

Congress has now joined in. On May 22, House Oversight Chairman James Comer sent document demands to both Kalshi and Polymarket and opened a formal insider-trading probe, asking how the platforms verify users and their locations, how they stop people routing around US rules through the offshore exchange, and whether the framework can cope with markets whose underlying asset is a geopolitical event. That the letters went to both the offshore-linked platform and the CFTC-regulated one signals Congress is treating this as an industry problem, not a single bad actor. So a second judgment sits alongside the first, also at high confidence. Official attention is widening on every axis at once, regulatory, congressional and criminal, with criminal enforcement clearly expanding, and the public record of indictments, lawsuits, subpoenas and rulemaking establishes it directly.

The Money Laundering Angle

Crypto-native platforms carry the same anti-money-laundering weaknesses as the rest of decentralized finance, and Polymarket’s offshore exchange is the obvious example. It has historically used permission-less onboarding, a Web3 wallet and you are in, with identity checks reserved largely for high-volume accounts. The Treasury’s Illicit Finance Risk Assessment of decentralized finance named Polymarket specifically, and noted that many such services subject to the Bank Secrecy Act do not comply with their AML obligations. Add pseudonymous transactions, mixers, chain-hopping and VPN access, and the controls thin out fast.

The appeal for laundering is the cover story. Deposit crypto, spread bets across a range of contracts, withdraw what looks like legitimate trading profit. The binary structure, plus the ability to hold both sides of a market across several accounts, makes wash trading and layering easy to run and hard to spot without serious on-chain analytics. The tracing firms, Chainalysis, TRM Labs, Elliptic, can follow it, but only after the transaction has cleared, and on a permission-less rail that is most of the battle lost already.

Confidence level: moderate: That prediction markets are being used to launder money is likely; the structural vulnerability is well documented, and Treasury has named it, but open sources cannot show how much of the activity is actually illicit as opposed to merely pseudonymous. So this rests on capability and incentive more than on confirmed cases, and is best treated as a moderate-confidence call.

War, Death, and the Moral-Hazard Problem

The contracts that should alarm a national-security reader most are those tied to military operations and the lives of political leaders, because they fuse financial speculation with operational risk. During the US-Israeli campaign against Iran in February and March, traders put more than $1 billion on essentially every dimension of the war, including strike timing, regime-change odds, casualty thresholds, and who would survive. Most of it ran through Polymarket’s offshore exchange, outside the reach of US commodity law, with the Khamenei market on Kalshi as the domestic counterpart.

The deeper danger is not the insider who trades on a coming operation but the insider who has a financial reason to want the operation to happen in the first place. If the people in the room deciding whether to strike are holding positions that pay out on a strike, a financial motive has been wired into a decision that is supposed to rest on national-security grounds and nothing else. 

Representative Greg Casar put the objection plainly enough, saying the country should not be in a situation where someone weighing whether to bomb or invade has hundreds of thousands of dollars riding on the answer. That was a hypothetical when he said it. The Van Dyke indictment, which alleges an insider betting on his own operation, is not.

What Congress is Trying to Do About it

Four bills are in play, and they aim at different parts of the problem:

• DEATH BETS Act (Schiff/Levin, March 10) would amend the Commodity Exchange Act to forbid any CFTC-registered entity from listing contracts touching terrorism, assassination, war, or an individual’s death, including anything that correlates closely to a death.
• BETS OFF Act (Murphy/Casar, March 17) goes further, banning wagers on government actions, terrorism, war, and assassination, choking off the payment rails offshore platforms rely on, and attaching criminal penalties for running such operations domestically.
• Prediction Markets Are Gambling Act (Schiff/Curtis, March 24, bipartisan) takes aim at the sports and casino-style contracts rather than the violence, barring registered entities from listing them.
• Prediction Markets Security and Integrity Act (Blumenthal) targets insider trading directly, adds customer protections, and would reverse the CFTC’s claimed preemption of state gambling law, which is the live legal question everywhere else in this memo.

The politics here are not especially partisan. Data for Progress polling has 61% of independents and 57% of Republicans behind a ban on wagering on government actions, and roughly 80% of voters opposed to markets on terrorism or assassination. The NFL’s Chief Compliance Officer has written to platform operators asking them to pull contracts the league finds objectionable, which is a notable step from a potential commercial partner.

How AI Changes the Calculus

AI runs through this problem on three tracks at once: traders use it for speed, platforms for surveillance, bad actors to manipulate prices and to hide. The first is already happening at scale and well evidenced; the third is mostly anticipated and not yet observed. The two should be kept apart.

Observed AI Activity

Autonomous trading agents have gone from concept to live deployment. Polystrat, launched on Polymarket in February by Valory AG through its Olas protocol, trades on its own with no human in the loop, In its first month it ran more than 4,200 trades and posted returns as high as 376% on single positions. By mid-March, 14 of Polymarket’s 20 most profitable accounts were bots, not human traders. One wallet, 0x8dxd, reportedly turned about $300 into more than $400,000 in a month, not by forecasting better but by exploiting latency, the lag between Polymarket’s prices and live feeds from Binance and Coinbase. A separate fully automated bot ran 8,894 trades on short-term crypto contracts and netted close to $150,000 catching the milliseconds when the combined price of “Yes” and “No” dipped below a dollar, windows no human trader can act on.

This corrodes the one thing prediction markets were sold on. If most of the volume and most of the price formation come from competing algorithms, the price stops being a crowd’s aggregated judgment and becomes the output of a machine-learning arms race. That is roughly what happened to equity and forex markets over twenty years of algorithmic and high-frequency trading, but those markets never claimed to be measuring human insight about the world. Prediction markets did, and journalists and policymakers cite their prices on that premise. 

Bot dominance also makes manipulation easier: systems watching prices and correlated feeds together can move faster than any surveillance can flag them, wash trading across bot-controlled wallets is trivial to run and hard to attribute, and on permission-less infrastructure the attribution problem is severe.

Anticipated AI Threats

The threat that warrants the most weight going forward is not the trading bot but synthetic media. Generative tools have lowered the cost of a convincing fake, and the play is obvious: fabricate a plausible signal about an event (a deepfake of a candidate withdrawing, a cloned audio clip of a CEO on unreleased earnings, a fake report of a strike), move the price, and exit before the debunking lands. 

On voice cloning, researchers report that short audio samples can now produce clones listeners struggle to distinguish from real recordings under test conditions; how well that holds against an alert or expert listener in the wild is less established and should not be overstated. The World Economic Forum’s 2026 Global Risks Report ranks misinformation at the top of its near-term systemic-risk list. 

Prediction markets sit at a bad intersection of it, both a target for manipulated information through their prices and an amplifier of it, since a manipulated price reported as a probability is itself disinformation. No confirmed, attributed case of this has surfaced yet; the concern is a capability and incentive both clearly present, not a documented event.

Confidence level: moderate: It is assessed as likely, at moderate confidence, that attempts to use synthetic media to move prediction market prices will increase over the next year or two; the technology, the incentive and the thin surveillance are all in place. What cannot be called is the timing, the frequency, or how often such attempts will work rather than simply be tried.

The Defensive Picture

Defense exists but appears to be losing ground. The analytics firms can cluster wallets and trace flows, and Polymarket says it runs a multi-layered monitoring system with outside partners, but on-chain surveillance is inherently after-the-fact on a permission-less rail, and the CFTC has not shown it can monitor this in real time at the scale now required. 

The common claim that offence is outpacing defense is plausible and consistent with what is observable, but hard to prove, and should be read as an informed assessment, not a measured fact. Clearer is the asymmetry of consequences, which is partly why the federal prosecutions matter so much. Felony exposure and seven-figure forfeiture deter in a way a $20,000 platform fine never will, and for now they are the only real deterrent in the system.

The Regulatory War

The Federal Posture

The federal posture shifted with the change of administration. Under Biden the CFTC took the view that election contracts could not be listed because they involved unlawful gaming; a district court rejected that, and after January 2025 the agency reversed itself completely. Chairman Michael Selig now asserts exclusive federal jurisdiction over event contracts and has set the agency up as the sector’s facilitator, no longer its skeptic. He scrapped the old restrictive proposal in January, withdrew the prior proposals in February, and ran an Advance Notice of Proposed Rulemaking that closed on April 30 with 3,534 comments.

On May 26, the agency sent its actual proposed rule, the first comprehensive federal rulebook for prediction markets ever drafted, to the White House Office of Information and Regulatory Affairs for review. The text is confidential for now and still has to clear public comment, so a final rule is months away at the earliest. 

The timing is noteworthy. It arrived the same day President Trump went on Truth Social to back the CFTC’s exclusive authority and promise to codify a permanent crypto market structure. Taken with the lawsuits against the states, the rulemaking and the endorsement look less like three separate developments than one unified effort to settle the jurisdictional question on federal terms before the courts do.

The State Counterattack

State governments have continued to challenge that position. More than a dozen have issued cease-and-desist orders, filed suit, or brought criminal charges on the theory that the platforms are unlicensed sportsbooks; the American Gaming Association puts lost state tax revenue at over $600 million. The CFTC’s response has been to sue the states first. Starting in April, with the DOJ alongside, it went after Arizona, Connecticut, and Illinois, then added New York, Wisconsin, and Minnesota: six in all. The six states targeted so far all have Democratic attorneys general; the agency’s engagement with Republican-led states has been limited to an amicus brief in Ohio. In Arizona a federal court has already paused the state’s criminal case against Kalshi, the judge signaling the agency will likely win on preemption.

Minnesota accelerated the legal dispute. On May 18, it became the first state to ban prediction markets outright, the prohibition tucked into a broader public-safety bill that passed both chambers comfortably. Operating, hosting, or advertising a market becomes a felony carrying up to five years and a $10,000 fine, though bettors are not criminalized and a late amendment carved out weather contracts. The CFTC sued the next day to block the law before its August 1 effective date. Nevada remains the only state with a court-ordered ban actually in force, after a state court extended a temporary block on Kalshi’s sports contracts, and Utah has banned proposition betting in language its sponsors say reaches these platforms.

The Circuit Split

The appellate courts are split, which is the main reason none of this is resolved. On April 6, a divided Third Circuit panel held in KalshiEX v. Flaherty that the Commodity Exchange Act preempts state gambling law as applied to sports event contracts on a registered DCM. This was the first federal appellate ruling on the question and a real win for the platforms; the majority called the contracts “swaps” and found both field and conflict preemption, the dissent argued gambling has always been the state’s business. The Ninth Circuit, which heard consolidated arguments in mid-April involving Kalshi, Robinhood, and Crypto.com, has been less receptive and declined to block Nevada. A split between circuits usually ends up at the Supreme Court, and most people watching expect it to, though probably not before 2027.

Confidence level: low: The ultimate outcome cannot be forecast with any confidence. Whether the Supreme Court takes a case, when, and how it rules on federal preemption are all open; the circuits are split, the controlling federal rule is unwritten, and the underlying legal questions are unsettled. Any confident prediction about the endpoint is guesswork, and this assessment holds the question at low confidence.

One point gets lost in the domestic noise: this entire fight, the CFTC, the circuit split, the state bans, is an argument among American institutions over which American authority gets to regulate. Many jurisdictions outside the United States have instead focused on restricting or prohibiting platform access.

The Rest of the World is not Waiting for Washington

The American debate hinges on a fine distinction: is an event contract a federally regulated derivative or a state-regulated bet? Most other jurisdictions decline to entertain the distinction at all. Their position, by now near-identical from Madrid to Delhi, is that a wager on an uncertain outcome is gambling regardless of what sits behind it on a blockchain, and an operator without a local gambling license is operating illegally. This has resulted in increasing restrictions on platform operations. By mid-2026, Polymarket faced restrictions in more than 33 jurisdictions, and its market outside the United States is contracting faster than the platforms can open new ground.

Europe’s Enforcement

Europe has been the most aggressive enforcement zone, partly because most European states already run mature gambling-licensing regimes and did not need new legislation to act. France’s gaming authority moved early, forcing Polymarket into a view-only mode for French users. The Netherlands escalated in February, Belgium made a referral in March, and Hungary and Portugal issued outright blocks back in January (the Portuguese action coming after roughly $120 million ran through the platform on that country’s presidential election). Spain is the most recent and the most explicit. On May 26 its gambling regulator published sanction proceedings in the official gazette and ordered ISPs to block both Polymarket and Kalshi, stating flatly that using crypto does not turn a wagering product into something else. That order caught Kalshi as well, and that is the telling part. The financial-venue framing at the centre of Kalshi’s entire US strategy carries no weight at all with European regulators.

A slower but more consequential European pressure is regulatory, not enforcement-driven. The grandfathering period under the EU’s Markets in Crypto-Assets regulation ends in July 2026, after which a crypto-settled platform serving EU users will need a Crypto-Asset Service Provider licence. None of the major operators hold one. MiCA was not written for prediction markets and does not cleanly classify them, leaving a gray zone that member states have filled with their own gambling law, but the licensing cliff is real and close. Separately, the EU’s DAC8 rules now feed automatic crypto-wallet data to tax authorities, making quiet retail use far more visible than a year ago. The UK, outside all this, simply treats fixed-odds prediction markets as gambling regardless of the technology; Polymarket is unlicensed there and blocks UK addresses.

Asia’s Response

Asia has produced the single sharpest action. India brought its Promotion and Regulation of Online Gaming Act into force on 1 May, reclassifying prediction markets as prohibited “money games,” then issued ISP-level blocking orders against Polymarket on May 21, with Kalshi following days later. India also went further than most Western regulators by warning VPN providers against enabling access, an attempt to close the usual workaround and not just the front door. Indonesia blocked Polymarket on  May 25, calling it an online gambling site disguised as a prediction market; the trigger was a contract on whether the sitting president would resign early. Market activity levels help explain the regulatory attention. Indian Premier League cricket contracts had in some weeks reached nearly half the volume of US baseball, and a single match in early May drew over $27 million.

The Platforms’ Approach

Rather than withdraw, Polymarket and Kalshi kept onboarding users in India for weeks after the ban took effect, and Polymarket has opened a formal lobbying effort in Japan, pushing into new ground as old ground closes. The approach resembles strategies previously employed by other rapidly expanding technology platforms operating in regulatory gray areas: enter the high-growth market, absorb the regulatory heat, keep operating until someone physically makes you stop. The long-term sustainability of that approach remains uncertain, and through 2026 enforcement has mostly been catching up with it.

Implications

Two implications matter. First, the convergence is itself a signal: when countries as different as Spain, India, and Indonesia independently reach the same legal conclusion within weeks, the “it’s a derivative, not gambling” argument is plainly losing abroad even as it gains ground in US federal court, and a platform’s legal status now depends heavily on which side of a border its user sits. 

Second, and less comfortable: blocking a platform does not make its users disappear, it pushes them toward VPNs, the permission-less offshore exchange, and whatever jurisdictions stay open, precisely the channels hardest to monitor. The wave of bans, whatever its consumer-protection merits, tends to worsen the illicit-finance and manipulation problems, not ease them. Fragmenting a market scatters the risk; it does not remove it.

Threat Actor Assessment

The threats above cut across several actor types, and separating them helps, because the access, the motive, and the right countermeasure differ in each case. The categories below run roughly from the most consequential for national security to the most diffuse.

Government and Military Insiders

These actors hold classified, sensitive, or pre-decisional information that confers a decisive advantage in markets tied to military operations, sanctions, and diplomacy. Their access is the highest-value and the hardest to detect, since the same secrecy that makes the information valuable shields the trade. The alleged conduct in the Van Dyke case is the template, a direct path from operational knowledge to a payout. The risk is not only the leak but the moral hazard: a financial stake in an operational outcome held by someone positioned to influence it.

Corporate Insiders

Employees with material non-public information on earnings, mergers, product launches, regulatory decisions, or internal metrics may use prediction markets as an alternative channel to exploit it, especially as contracts increasingly track public-company events. The Spagnuolo allegations show how internal data (in that instance search trends and not classic financial information) can be leveraged in markets tied to observable outcomes. Most corporate compliance regimes do not yet treat this as covered conduct, which widens the exposure.

Organized Crime and Laundering Networks

Crypto-native infrastructure offers these actors layering, wash trading and the movement of funds under the appearance of legitimate trading. Pseudonymous wallets, offshore platforms, and cross-chain movement complicate attribution and enforcement. Their interest is less informational edge than the laundering utility of a venue that produces plausible profit-and-loss records on demand.

Foreign Intelligence Services

A capable state actor can use prediction markets two ways at once. As collection: anomalous trading around military operations, elections, sanctions, or diplomacy can offer a read on what other informed participants expect. As influence: coordinated activity can push a market price, and therefore the widely cited probability it represents, to shape public or elite perceptions of how likely an event is. The same permission-less, cross-border infrastructure that frustrates criminal attribution frustrates it here too.

Automated and AI-Enabled Actors

Autonomous systems execute at a speed and scale no human can match, which suits latency arbitrage, coordinated activity across wallet clusters, and rapid manipulation. The category overlaps the others rather than standing apart: a criminal network, a state actor, or an opportunistic firm may all operate through automated agents, and generative AI lowers the cost of synthetic content paired with a trading strategy. The distinguishing feature is tempo, the actor can act and unwind faster than current surveillance can flag it.

Intelligence Implications

For a collection shop, the implication is straightforward. Prediction markets are not only platforms to be regulated but a place to look for indications and warning. An analyst on a military, geopolitical, or economic problem can fold market activity into the wider picture, both as an early read on what informed participants expect and as a surface a hostile actor may be trying to bend. The same data serves both purposes, which is why it has to be read with care, not taken at face value.

Potential indicators warranting monitoring include:

• Significant volume increases immediately preceding major geopolitical or military events.
• Concentrated positions taken by newly created or previously dormant accounts ahead of a discrete outcome.
• Coordinated trading across clusters of wallets or otherwise related entities.
• Sharp market moves tracking to unverified media reports or content with signs of synthetic origin.
• Repeated, improbably accurate trading by individuals with privileged access to relevant information.
• Cross-platform activity that suggests an effort to move both market pricing and the surrounding public narrative at once.

None of these is conclusive alone, and each has innocent explanations (sophisticated retail traders, legitimate hedging, ordinary news reaction). Their value is as a screen: a cluster around a sensitive event is a reason to look harder, not a finding in itself.

Where this Goes

Earlier assessments predicted that a first federal insider-trading prosecution was probable within a year, and has already been overtaken; two are in the courts. The question now is convictions, and a guilty verdict or a serious plea in either Van Dyke or Spagnuolo would likely trigger a second wave of charges around election and Iran trades. That follow-on wave is assessed as more likely than not within twelve months, at moderate confidence.

The jurisdictional fight will not resolve cleanly or quickly. The CFTC rule has months of process ahead of it, the six state suits are the near-term tell, and the Third-versus-Ninth Circuit divergence points toward a Supreme Court that may not rule before 2027. If federal preemption wins, the platforms get nationwide access under CFTC oversight; if the states win, the US market fragments into something like state-by-state sports betting. Neither outcome can be called with confidence. 

Abroad the direction is clearer and runs the opposite way, enough to support a moderate-confidence judgment: the MiCA cliff in July, the spreading “it’s gambling” consensus, and regulators’ willingness to chase VPNs all point to a narrowing global footprint. 

The likely shape, then, is a sector more firmly legitimized inside the United States while walled out of much of everywhere else: a strange position for a business whose offshore exchange was built on global reach. And if Congress moves on war, death or sports contracts, even the US product falls back on economic indicators, weather and corporate milestones, the lowest-friction categories anywhere. That retreat, chosen or forced, is the most probable medium-term outcome.

Illicit use is likely to keep climbing before it falls off. Rapid growth, crypto-native offshore venues, and a still-thin record of completed enforcement all favor abuse, even as the first prosecutions begin to bite. The actor categories outlined in the threat assessment warrant continued monitoring, with one addition the categories do not capture: match-fixing and event manipulation in sports, entertainment, and content-creation markets, where a single participant can plausibly move the underlying result.