United States v. Van Dyke: The Polymarket Insider Trading Case and the First Criminal Test of the Swap Definition
By R Tamara de Silva | De Silva Law Offices, LLC | August 27, 2026
The litigation over prediction markets keeps returning to two questions. The first is definitional: is an event contract a "swap" under the Commodity Exchange Act? The second is preemption: if it is, how much room does federal law leave the states? The state cases involve both, in different combinations; a Connecticut court ruled against Kalshi on both this month, at the preliminary injunction stage and on independent grounds. The criminal case now before Judge Garnett in Manhattan is different. The government has charged Gannon Van Dyke, an Army Special Forces master sergeant, with insider trading, alleging that he used advance knowledge of the operation that removed Nicolás Maduro to buy Polymarket contracts on Maduro's ouster and cleared more than $400,000. That case presents the definitional question alone, with no preemption overlay, and it presents it under rules of construction that give the accused the benefit of ambiguity.
On August 24, Judge Margaret M. Garnett of the Southern District of New York granted the CFTC leave to file an amicus brief, a friend-of-the-court submission, opposing Van Dyke’s motion to dismiss the first federal insider trading prosecution involving event contracts.¹ The defense had objected in unusually direct terms. Its letter called the agency a “regulatory wolf” seeking to pile on as a “second prosecutor.” Judge Garnett accepted the brief anyway, noting that she would give it “its appropriate weight,” and allowed both sides until September 9 to respond in no more than ten pages. With that order, the principal papers are all before the court. The defense memorandum, the government’s opposition, and the agency’s own account of its statute now sit before a single district judge, with trial set for December.
When the indictment was unsealed in April, this firm wrote that the government had chosen the cleanest fact pattern available for its first criminal prediction market case https://www.desilvalawoffices.com/articles/blog/2026/april/the-first-criminal-polymarket-insider-trading-ca/. Clean, because everything a prosecutor needs was in one place. The defendant was identified. He had signed nondisclosure agreements days before the trades. He is alleged to have helped plan the operation he traded on, to have used an account in his own name, and to have turned roughly $34,000 into more than $400,000 when the contracts resolved. The chain from duty to information to trade to profit required no imagination.
We also noted what the government had not charged: the anonymous accounts that, according to contemporaneous reporting, placed remarkably timed positions ahead of American military action, including roughly $3 billion in oil futures then under CFTC investigation and a cluster of Polymarket accounts that profited ahead of the Iran strikes. Those cases involve bigger money, unnamed traders, and harder problems of attribution and jurisdiction. They remained uncharged then; nothing public suggests that has changed.
Four months later, the facts have produced a hard case anyway. Not about the conduct, but about the instrument. Whether Van Dyke misused information he was bound to protect is the kind of question juries decide. Whether the contracts he is alleged to have traded are "swaps" within the CFTC's jurisdiction is the question dividing courts across the country, and it comes first. The motion, which takes the indictment's allegations as true for its own purposes, puts that question squarely before Judge Garnett. Van Dyke has pleaded not guilty. If the Polymarket contracts were not swaps, three of the five counts fall, and the reasoning would carry well past this docket.
Why the Defense Says Polymarket Event Contracts Are Not Swaps
The defense memorandum opens with a line built to be quoted: criminal courts are not laboratories where prosecutors test new ideas about whether conduct is criminal.
The defense's statutory argument runs through Section 1a(47)(A)(ii) of the Commodity Exchange Act, the statute the CFTC administers. This section defines a swap to include a contract providing for payment dependent on an event or contingency "associated with a potential financial, economic, or commercial consequence." On the defense's reading, Dodd-Frank's definition covers instruments designed to hedge price risk and facilitate price discovery. A wager on whether Nicolás Maduro would be out of power by January 31, according to the Van Dyke defense, is neither.
The memorandum grounds this reading in recent judicial authority. Three courts in the state prediction market litigation have adopted versions of it. The Ohio court read the definition to reach events that are "inherently economic" or that traditionally and directly affect commodity prices, and the defense adopts that formulation as its own. Judge Oliver in Connecticut concluded this month, at the preliminary injunction stage, that Kalshi's sports contracts are unlikely to qualify as swaps at all. The Nevada court spoke to Dodd-Frank's purpose: Congress was bringing risky financial products out of the shadows after 2008, it wrote, not "enabling nationwide gambling on CFTC-designated exchanges."
None of those rulings binds Judge Garnett, and all of them arose in civil postures. They serve two functions in the memorandum nonetheless. They establish that the narrow reading has judicial support, and they lay the foundation for the fair-notice argument that follows: if federal courts continue to divide over what the definition means, the defense will argue, a defendant cannot fairly be charged with knowing what it covers.
The memorandum also makes a legislative-history argument, and it is more substantial than the press coverage suggests. Dodd-Frank's swap provisions were a response to the over-the-counter derivatives that destabilized the financial system in 2008. Nothing in the statute's record, the defense argues, suggests that Congress intended the same regime to reach wagering on world events, and a change of that magnitude is not one Congress makes silently. The strongest support comes from the regulators themselves. In 2012, the CFTC and SEC jointly adopted a rule further defining "swap," and in it they excluded insurance contracts, a category the bare text of subparagraph (ii) would appear to reach, on the ground that regulating insurance would not serve Title VII's purposes. The defense treats that exclusion as an admission: the agencies themselves read the definition as limited by its purpose, and the government's reading in this case has no limit at all.
The memorandum's second statutory argument is directed at a different element of the definition. The dispute discussed above concerns the nature of the event: whether Maduro's removal was "associated with" a financial, economic, or commercial consequence. This argument assumes that it was and contends that the counts fail regardless, under the clause that precedes it. The statute requires that a swap's payment be "dependent on the occurrence, nonoccurrence, or the extent of the occurrence" of the event. Payment, that is, must turn on whether the event in fact occurred. The defense treats dependence as a separate element, and it is the element, on the defense's account, that these contracts cannot satisfy.
Dependence, on the defense's reading, requires a mechanical link between event and payment, what the memorandum calls a causal architecture. The event occurs, a number moves, and the number determines who is paid, with no discretion anywhere in the chain: an index, a price everyone can observe on a marketplace, or official data. Conventional swaps work this way. A weather swap settles on temperature readings from the National Weather Service, collected by an independent processor with no stake in the result. An interest-rate swap settles on the published Secured Overnight Financing Rate. A commodity swap settles on an exchange price. In each case, the trigger is external to, and independent of, the parties who declare the result.
The Polymarket contracts, the defense argues, depart from that structure at the point that matters. The contracts settled on "a consensus of credible reporting" about the scenarios they posited. When a result was disputed, the platform's rules made the collective judgment of market participants, participants with a financial stake in the answer, the operative trigger of payment. Settlement, in other words, runs through the deliberation of interested parties rather than through measurement. Congress did not define swap to include "wagers on what a group of self-interested bettors will later agree happened." It defined the term to include payments keyed to whether an event did or did not, in fact, occur. If the causal link is absent, the counts fail regardless of how the consequence question is resolved.
The definition also contains a catch-all. Subparagraph (iv) covers any contract that "is, or in the future becomes, commonly known to the trade as a swap," and the government's papers invoke it. Even if the contracts satisfied no other part of the definition, the argument runs, they would qualify because the trade now calls them swaps.
The defense's answer begins with a question: known to what trade, and since when? For as long as such contracts have existed, the defense contends, they have been known as gambling. The swap characterization is recent, and it originated with the platforms, which had a commercial reason to adopt it. Call the product a swap and it falls within the CFTC's exclusive jurisdiction; if it falls within the CFTC's jurisdiction, state gambling law is preempted. The characterization, in other words, is the platforms' preemption defense. A label the seller adopted to secure a jurisdictional advantage, the defense argues, is not what the trade commonly knows. The defense adds a constitutional limit: due process does not permit a court to expand a criminal statute retroactively, to reach conduct that was not understood to be covered when it occurred.
The defense's final statutory argument rests on Loper Bright. Start with what the provisions actually say. Section 9(1) prohibits manipulative and deceptive devices. Rule 180.1 tracks that language. Neither one mentions insider trading. The view that they nonetheless reach it is the CFTC's interpretation of its own statute, and under Loper Bright, the Supreme Court's 2024 decision ending judicial deference to agency interpretations, that view is worth only whatever persuasive force it earns. A court owes it nothing. The point carries extra weight here, the defense argues, because the statute is criminal, and courts do not defer to the government's interpretation of laws that carry prison time. Nor has case law filled the gap. The insider trading theory has produced guilty pleas in one Texas prosecution and one appellate affirmance, the Fifth Circuit decision the government now relies on.
The constitutional arguments build on the statutory ones. The first is fair notice. Due process requires that a criminal statute give a person of ordinary intelligence fair warning of what it prohibits, and the defense contends that this one cannot. The record it assembles is the state of the field itself. The federal courts hearing the state prediction market cases are divided. Congress has introduced a dozen bills addressing whether these contracts are swaps. The CFTC and SEC have asked the public for comment on where the definitional lines sit. If the institutions that write, enforce, and interpret the statute cannot agree on what it covers, the defense argues, an individual trader cannot be charged with knowing.
The wire fraud count turns on a different element. Wire fraud requires a scheme to obtain money or property, so the count stands or falls on whether classified military planning information is "property" in the government's hands. The defense contends that it is not, and it grounds the argument in the Second Circuit's decision in Blaszczak. There, the court held that confidential government information about a regulator's plans was not property, and on remand the government itself told the court that confidential government information typically must have economic value in the government's hands to qualify. Military plans, the defense argues, fail that test. The government does not sell or license them; they are an instrument of sovereign power, not an asset held for commercial value. Lenity underlies all of it. If, after every tool of construction, a criminal statute remains grievously ambiguous, the ambiguity is resolved in the defendant's favor, and the defense invokes that rule across each of its statutory arguments.
How the Government and the CFTC Amicus Defend the Swap Definition
The government's opposition begins with the theory of the case, which it says is anything but novel. On the government's account, Van Dyke took information he had pledged to keep confidential and, in breach of that duty, used it for personal profit; the platform is new, the crime is not. The opposition then takes up the swap definition element by element. Section 1a(47)(A)(ii) imposes four requirements, and in the government's telling, Van Dyke disputes only the fourth: economic consequence. The amicus makes the same point from a different direction. The defense does not deny that Maduro's removal was an event; its arguments go to consequence and to how the contracts settled.
The government's answer on consequence is empirical rather than abstract. Within a day of the raid, its papers recount, Venezuelan sovereign bonds rose roughly a third, and bonds of the state oil company rose more than forty percent. The intervention has since produced, again by the government's account, a $100 billion reconstruction program and more than $13 billion in United States receipts from Venezuelan oil. The statute requires only a potential consequence. On this record, the government argues, the consequences were actual.
The amicus brief contains the agency's fullest public account of its own definition, and its textual points are direct. The statute says "potential" consequence, and potential is a lower bar than actual: a contract can qualify even if the consequences never materialize, because what might happen always exceeds what does. The defense's "inherently economic" gloss then collides with the statute's own list. Subparagraph (iii) names weather swaps and emissions swaps as swaps, and neither weather nor emissions is inherently economic. A limiting principle that would read out contracts Congress included by name, the amicus argues, cannot be the right one.
Courts read statutes on the presumption that Congress chose every word for a reason. An interpretation that leaves part of a statute with no work to do, what lawyers call surplusage, is therefore disfavored. Both sides accuse the other of using surplusage.
The defense's version argues that if the event clause in subparagraph (ii) reaches any contract tied to a consequential event, then the specific instruments Congress listed elsewhere in the definition were unnecessary, and a reading that makes them unnecessary should be rejected. The amicus answers that the defense's reading has the same defect in reverse. If the event clause covers only ground the neighboring provisions already cover, then the event clause is the one left with nothing to do. The Supreme Court addressed exactly this stalemate last year in Bufkin v. Collins: when both readings produce some redundancy, the canon cannot choose between them, and it drops out of the analysis.
The amicus then offers a second route to the same destination, one that has received little attention. The Act's definition of "commodity" is not limited to wheat and oil. Under Section 1a(19)(iv), it also includes an occurrence beyond the parties' control that carries economic consequence, what the statute calls an excluded commodity. If that is right, Maduro's removal was itself a commodity. The Polymarket contracts were binary options on that commodity, and an option on a commodity falls within the swap definition through its first subparagraph. The contracts would qualify, in other words, even on the defense's own commodity-centered reading.
The amicus also invokes the Special Rule, the Dodd-Frank provision that permits the CFTC to bar certain event contracts as contrary to the public interest. The provision expressly names contracts involving "war" among those the agency may review. Congress, the amicus argues, thereby presupposed that war-related event contracts fall within the agency's jurisdiction; on the defense's reading, the provision would govern an empty category. The argument has administrative history behind it. The CFTC's 2008 concept release recorded the agency's regulation of event contracts on the declaration of war, presidential elections, and world population levels. Its 2012 order prohibiting a registered exchange's political event contracts was itself an exercise of Special Rule authority, over products the agency treated as within its jurisdiction.
The government also draws a boundary of its own, and it is the boundary the state cases have been contesting. The sports rulings the defense relies on, the opposition argues, answered a narrower question. A scheduled game will take place; the only uncertainty is its outcome, and those courts asked whether an outcome is a separate "event" under the statute. The Maduro contracts presented no such puzzle. They asked whether an event would occur at all, which the government describes as the paradigm case, one that qualifies under even the narrowest construction any court has adopted. The distinction separates economically consequential occurrences from game outcomes, and here it is drawn by the prosecution. It is the same boundary the Washington injunction and the CFTC's emergency order converged on two weeks ago, and the same two-tier structure this firm has urged in its comment letters to the CFTC.
On the insider trading theory itself, the government's answer is history. Section 9(1) was modeled on Section 10(b) of the Securities Exchange Act. Under Section 10(b), the misappropriation theory imposes liability on a person who trades on confidential information taken in breach of a duty, and the Supreme Court settled that theory in O'Hagan in 1997. When Congress borrows statutory language, the settled judicial meaning comes with it. The amicus adds the legislative record. The provision's Senate sponsor said on the floor that the words were taken from Section 10(b) precisely so that courts would read the new authority through seventy-five years of accumulated case law.
Rule 180.1 enforcement is not new either. The CFTC has brought misappropriation cases under the rule for a decade, and last year the Fifth Circuit affirmed a criminal conviction under Section 9(1) and Rule 180.1, applying Section 10(b) precedent to do it. The papers also revisit Section 4c(a), the companion prohibition aimed at trading on confidential government information. Congress added it after then-Chairman Gary Gensler testified that the CEA, as written, would not reach the scheme depicted in Trading Places, which is why the provision carries Eddie Murphy's name. We walked through Rule 180.1 and the Section 4c(a) prohibitions in April. The doctrine has not changed, and the government's papers deploy it exactly as expected.
On the question of purpose, the amicus does not concede the defense's premise that derivatives exist to hedge. The Second Circuit rejected that idea decades ago: a derivatives market could not function if only hedgers traded, because someone must take the other side. Speculators who assume the risk that hedgers shed are part of the market's design, and the CEA protects them. The amicus also contests the premise as a factual matter. Institutional block trading in prediction markets launched this month. A sports platform uses event contracts to manage its own risk. And a Manhattan bar hedged its free-drinks-if-the-Knicks-win promotion on Kalshi. Hedging in these markets, the brief argues, is no longer hypothetical.
On price discovery, the defense draws a distinction: conventional derivatives discover prices, while prediction markets discover only probabilities, which the CEA was not built to protect. The amicus answers that the distinction does not hold, because in these markets the probability is the price. The $0.09 Van Dyke is alleged to have paid on December 30 was the market's price for a dollar contingent on Maduro's capture by January 31, which is to say the market's estimate that capture was a nine percent probability. A weather swap's price is a probability in the same way. The amicus also responds to the defense's reliance on unenacted legislation. Bills that never passed prove little in either direction, it argues, but Congress has not been silent. When the Senate barred its own members from trading on prediction markets earlier this year, the text of its rules change described those transactions as swaps as defined in Section 1a of the CEA.
The amicus's answer to the settlement argument matters beyond this case, and it begins with the defense's own example. The defense offered the Latin American Dated Brent strip as the model of a proper swap. Dated Brent is a benchmark price for physical crude oil cargoes, and the amicus observes that the number does not generate itself. A price-reporting agency surveys market participants, applies its own analysis, and publishes an assessment. The model swap, in other words, settles on a humanly constructed figure.
The amicus's other examples run the same direction. The London gold auction produces a reference price from orders placed by dealers with positions of their own. Credit default swaps settle only after a committee of banks and funds that trade those instruments determines whether a default occurred, with an auction fixing the payout. If resolution by interested parties disqualified an instrument from swap status, the amicus argues, the credit default swap market could fail the same test.
The Washington injunction showed that contract taxonomy has become a jurisdictional fact. The briefing in this case adds settlement mechanics to the list. How a platform sources the facts that resolve its contracts, and how it handles disputes about them, is now an argument for or against federal jurisdiction. Platforms should design those mechanisms with Section 1a(47) in mind, and they should document the design.
Fair Notice and the CFTC’s 2022 Polymarket Order
The fair-notice question carries different weight in a criminal case, and the amicus builds a record for it that the civil cases have never needed. The record runs in three steps, each closer to the defendant than the last. The CFTC has regulated event contracts since at least 1992. In 2012, it prohibited political event contracts on a registered exchange, an order that treated such products as within its jurisdiction. And in January 2022, it entered a settled order against the operator of Polymarket itself. That order stated that the platform's event contracts, each composed of a pair of binary options, "constitute swaps under the CFTC's jurisdiction."² The platform on which Van Dyke is alleged to have traded was told, in a public order, four years before the trades at issue, that its products were swaps. The amicus adds one more contemporaneous fact: while Van Dyke was allegedly trading on Polymarket, a CFTC-registered exchange was listing Maduro contracts of its own.
The defense's response is that the swap characterization is recent, contested, and commercially convenient for the platforms that adopted it. Against subparagraph (iv), which asks what the trade commonly calls these contracts, that response has force. Against a public order naming the very platform, it has less. A trader can dispute what an industry believes. It is harder to dispute what the regulator told the platform, in writing, before the trades were made. That is why the 2022 Polymarket order may prove the most important single document in the notice fight.
Why a Criminal Case Raises the Stakes for the Swap Definition
The word "swap" is being litigated in nine civil actions and in multiple courts of appeals, with two emergency motions pending before Second Circuit panels. Every one of those disputes proceeds under civil standards: preliminary injunction factors and the ordinary tools of construction. This case proceeds under criminal ones. Lenity, fair notice, and vagueness doctrine share a common feature: each resolves doubt against the government. In the nationwide litigation over the definition, this is the only case where those rules apply.
The defense's strongest structural argument is the litigation map itself. Nine states, two federal agencies, and a dozen bills are contesting what the definition means, and the defense asks how a statute can be called clear while that contest runs. The government's answer narrows the frame. Disagreement at the margins, it argues, does not create ambiguity at the core, and no court anywhere has suggested that a contract on the capture of a head of state, one sitting on the world's largest oil reserves, falls outside a definition keyed to economic consequence. The standards give Judge Garnett an asymmetric task. She can deny the motion without deciding where the definition's outer edge lies. To grant it, she would have to conclude either that the text plainly excludes these contracts or that its ambiguity is grievous.
The stakes are asymmetric as well. A denial that reaches the merits would be the first square judicial holding that geopolitical event contracts are swaps; as the amicus notes, Judge Torres, in the same courthouse, assumed the point without deciding it in the New York Kalshi case. And a holding that survives lenity, fair notice, and vagueness would carry a weight no preliminary injunction ruling can, because it would mean the definition was clear enough to support a criminal conviction. A dismissal on lenity or fair-notice grounds would move just as fast in the other direction, into the states' briefs almost immediately. Either way, the ruling will travel.
One more feature of the briefing deserves attention: both sides are relying on the same Supreme Court decision. The defense invokes Loper Bright against Rule 180.1, arguing that the agency's reading of Section 9(1) receives no deference and must convince the court on its own merits. The CFTC is operating under the same decision from the other side, and its conduct shows it. The agency has stopped asking for deference. Its own civil case against Van Dyke was stayed on August 10, in this same courthouse. Two weeks later, the stayed plaintiff returned through the amicus door, over the defense's objection, admitted on the ground that its regulatory perspective could aid the court.
The stakes of that shift are easy to miss. CFTC Chairman Michael Selig recently put the global derivatives market at $1.2 quadrillion in notional value, nearly half of it within the CFTC's jurisdiction, and the amicus puts the volume of the event contract markets alone in the tens of billions of dollars. Before Loper Bright, an agency with that portfolio could expect courts to defer to its reading of its own statute. It no longer can. If the CFTC wants a court to adopt its view of the word "swap," it must do what any litigant does and persuade, and the amicus brief is that persuasion at work. The role is more modest than the one the defense letter feared. It may also prove more durable. An emergency order directs a single company and binds no court. An argument a court adopts becomes precedent, and precedent binds everyone.
What to Watch in United States v. Van Dyke
Supplemental filings responding to the amicus are due September 9. Van Dyke returns to court on September 28, and trial is set for December. The wire fraud count deserves watching in its own right. Blaszczak drew a line between the government's operational information and its regulatory information, and whether classified military planning falls on the property side of that line will matter to every future case in which government information is monetized. The government also has a second theory, that the counterparties' money satisfies the property element regardless of how the information is characterized. If that theory is accepted, the property question never has to be answered.
The center of gravity, though, is the swap definition. In April, we wrote that the government had picked the easiest case to charge. The briefing shows why even the easiest case was never going to be easy. Before a jury hears a word about nondisclosure agreements, a federal court must decide what a swap is, and it must decide under the rules of construction least forgiving to the government. One word is carrying the weight of an asset class. In this courtroom, for the first time, it carries it under the criminal law.
Endnotes:
¹ United States v. Van Dyke, No. 1:26-cr-00156 (S.D.N.Y.). The CFTC’s parallel civil action, CFTC v. Van Dyke, No. 1:26-cv-03369 (S.D.N.Y.), was stayed on August 10, 2026, pending the criminal case.
² In re Blockratize, Inc. d/b/a Polymarket.com, CFTC Dkt. No. 22-09 (Jan. 3, 2022).
De Silva Law Offices, LLC follows the prediction market space closely and writes about it regularly, from the state enforcement actions and the CFTC's emergency orders to the definitional questions now before the federal courts. The firm represents exchanges, intermediaries, market participants, and new entrants in the prediction market and event contract space, on matters including CFTC and NFA registration, compliance, and enforcement defense. Questions about this article or the firm's practice may be directed to info@desilvalawoffices.com.
NB: This article is for informational purposes only and does not constitute legal advice. Reading this article does not create an attorney-client relationship. Mr. Van Dyke is presumed innocent unless and until proven guilty.