Two bodies of law, one product
The legal fight over prediction markets is not about whether they are good or bad. It is about which of two bodies of law governs a contract on the outcome of a game, and the answer decides who licenses it, who taxes it, who can use it and who can stop it.
Gambling law in the United States is state law. Each state decides whether betting is legal, on what, through whom and at what tax rate, and the federal government's role has historically been to help states enforce their choices. Derivatives law is federal. The Commodity Exchange Act gives the Commodity Futures Trading Commission authority over futures, options and swaps, and it contains a clause stating that the commission's jurisdiction over futures and swaps traded on a designated exchange is exclusive. A prediction market that is a designated exchange says the second body of law applies and displaces the first. A state that licenses sportsbooks says a bet is a bet, and no federal registration changes what it is.
Everything in the litigation is a variation on that argument.
The federal framework
Designation. A designated contract market is an exchange that has satisfied the CFTC's core principles: financial resources, governance, surveillance, rules against manipulation, customer protection and reporting. Designation is expensive and slow (Kalshi's, granted in 2020, followed a nine-month review), which is why for years the main American event-contract venues were small academic projects such as the Iowa Electronic Markets, which the CFTC first allowed in 1992. The venues that grew in 2024 and 2025 either obtained designation or bought an exchange that already had it, as Polymarket did through QCX, an exchange designated in July 2025 that now operates as Polymarket US.
Event contracts. The act contemplates contracts based on the occurrence of an event and gives the commission a specific power over them. A provision added in 2010 allows the CFTC to prohibit an event contract if it involves terrorism, assassination, war, gaming or an activity unlawful under state or federal law, and if the commission finds the contract contrary to the public interest. The commission's rule implementing that power, Regulation 40.11, is the one the disputes have turned on, and in June 2026 the commission proposed amending it to define "involve" and "gaming" and to set out how it would judge the public interest contract by contract. "Gaming" is on the list. What gaming means, and whether the commission must act for the prohibition to bite, are the questions.
The election precedent. The first major test was not sport. In September 2023 the commission refused to allow an exchange to list contracts on which party would control Congress, reasoning that they involved gaming and activity unlawful under state law and were contrary to the public interest. The exchange sued. In September 2024 a federal court in Washington held that the commission had exceeded its authority: the contracts involved elections, which are neither "gaming" nor unlawful activity in the sense the statute used, so the commission had no power to carry out a public-interest review of them at all. An appeals court declined to block the ruling, election contracts traded through the 2024 presidential election, and in May 2025 the commission dropped its appeal. That sequence established that a designated exchange could list contracts on events that most states would not allow as bets, and it set the pattern for what followed.
Sport changes the stakes
In December 2024 one of Kalshi's competitors began listing contracts on the outcomes of sporting events, and Kalshi followed a month later, with contracts on games including the Super Bowl. Election contracts had annoyed regulators; sports contracts threatened an industry. By March 2025 gaming regulators in states with licensed sports betting, Nevada and New Jersey among them, had sent cease-and-desist letters asserting that the contracts were unlicensed sports wagering under state law. The exchanges did not comply. They sued.
The exchanges' argument has three parts. First, exclusive jurisdiction: the act says the CFTC's authority over swaps and futures traded on a designated exchange is exclusive, a sports event contract is a swap, and a state law that would prohibit those contracts is preempted. Second, the gaming provision itself: Congress gave the CFTC, not the states, the power to prohibit event contracts that involve gaming, so the existence of that power shows Congress intended federal law to occupy the field. Third, uniformity: a national exchange cannot operate under fifty different rulebooks, which is precisely why Congress created a single federal regulator for derivatives.
The states' argument also has three parts. First, the gaming provision proves the opposite of what the exchanges say: Congress listed gaming and activity unlawful under state law as things the commission may prohibit, which presupposes that state law continues to define what is unlawful. Second, police powers: the regulation of gambling is a core state function that federal statutes are not read to displace without a clear statement, and the act contains none. Third, the contract itself: a yes-or-no position on which team wins, paid at one dollar, is a sports bet in every respect that matters, and calling it a swap does not change what a customer is doing. On this reading the outcome of a game has no inherent financial consequence, so the contract falls outside the act's definition of a swap altogether.
Where the courts stand
At the time of writing, in September 2026, the litigation has produced a split.
Federal district courts in Nevada and New Jersey granted Kalshi preliminary injunctions in April 2025, accepting the preemption argument as likely to succeed and finding that the exchange would be irreparably harmed by state enforcement. Others refused: a Maryland court in August 2025, and the Nevada court itself, which later in 2025 dissolved its own injunction after concluding that sports contracts are not swaps. Later district decisions went both ways, for the exchange in Tennessee and Arizona and for the states in Ohio and New York. The appeals courts have now split too. In April 2026 the Third Circuit, by two votes to one, upheld the New Jersey injunction, holding that the exchange was likely to show its sports contracts are swaps within the CFTC's exclusive jurisdiction. In August 2026 the Ninth Circuit, unanimously, upheld the dissolution of the Nevada injunction, holding that the contracts are likely not swaps at all and that the CFTC's own rule bars gaming contracts. In September 2026 it held in a separate case that sports contracts entered into from tribal land were class III gaming under the Indian Gaming Regulatory Act, a statute that gives tribes and states, not the CFTC, control over that activity. New Jersey asked the Supreme Court to take the question on 2 September 2026 (No. 26-299; Kalshi's response is due 9 November 2026), and Robinhood filed its own petition against Nevada (No. 26-338). On 25 September 2026 the Sixth Circuit, in appeals from Ohio and Tennessee, also ruled against Kalshi, holding that its sports contracts are likely not swaps and that, even if they were, federal law would not preempt the states' gambling laws. A Fourth Circuit appeal from Maryland is still pending, so the Supreme Court is where the question sits.
The commission's own position shifted with the administration. Under one chair it fought election contracts; under its later leadership it dropped that fight, filed briefs supporting the exchanges against the states and, from April 2026, sued states directly, among them Arizona, Connecticut, Illinois, New York, Wisconsin and Minnesota, to block their enforcement. A regulator's litigating position is not law, but it tells the courts how the agency reads its own statute, and courts give that some weight.
A separate front is tribal. Tribes operate gaming under compacts that give them exclusivity in many states, and a contract on a game available to everyone on a phone competes with that exclusivity. Tribal governments have sued, and the class III ruling may prove the most consequential decision yet, because it does not depend on preemption of state law at all: the Ninth Circuit held that the act and the tribal gaming statute address different questions, so the CFTC's exclusive jurisdiction does not reach class III gaming on tribal land.
What each outcome would mean
If the exchanges win outright, a federally designated exchange may offer contracts on any event, including sport, to any adult in any state, subject only to CFTC rules, with no state licence, no state gaming tax and, on the largest exchange, a minimum age of 18. Licensed sportsbooks would be competing against a product that carries a fraction of their regulatory cost, and the rational response, which several have already taken, is to become an exchange or partner with one: FanDuel partnered with CME Group and DraftKings bought a designated exchange in 2025.
If the states win outright, a contract on a sporting event is sports wagering wherever a state says it is, and an exchange must obtain a state licence, or geofence the state out, to offer it. Election and economic contracts would likely survive on the election precedent, although Nevada's challenge to election contracts has been sent back to the district court. Sports contracts would revert to being available only where sports betting is legal, through licensed operators, in most states at 21.
The likeliest outcome may be neither. A Supreme Court ruling, a legislative fix or CFTC rulemaking that draws a line between contracts with a plausible hedging or informational purpose and contracts that are wagers on games would be a middle course, and the line-drawing would generate a decade of further argument. The CFTC has started down that road: on 9 October 2026 it proposed defining "swap" to include event contracts on sports, politics, culture and weather, and issued an interim final rule excluding casino-style gambling, including sportsbook wagers and casino games, from the definition.
Outside the United States
Almost everywhere else, the question has been answered quickly and in the same way. A contract that pays on the outcome of an uncertain event, offered to the public for money, is a bet, and offering it requires a gambling licence. Gambling regulators in Britain, France, Belgium, Singapore, Denmark, Lithuania and South Korea, among others, have said so, and several, including France in July 2026 and Denmark under a court order the same month (its July 2026 list of blocked sites includes polymarket.com), have ordered internet providers to block the largest crypto-native venue, Polymarket. Britain's regulator said in February 2026 that a prediction market launching there could not classify itself as a non-gambling product, that current products would appear to need a betting intermediary licence, and that the betting exchanges operating in Britain since 2000 are what they most resemble.
The crypto-native venues have argued, in various forms, that they are not offering gambling in the country concerned, as Polymarket did when South Korean regulators reviewed it. None of the regulators above has accepted that. The venues' practical response has mostly been to geoblock the country, as Polymarket did in France after the regulator's first formal notice, serve the rest of the world and pursue the American market through a designated exchange, which is the one place the "not gambling" argument has won anything.
What to take from this lesson
The legal question is jurisdictional, not moral. The exchanges hold a federal designation and a clause that says the CFTC's jurisdiction is exclusive; the states hold the definition of gambling and a provision that names gaming as something the commission may prohibit. District courts split, the appeals courts have split too, with the Third Circuit for the exchanges and the Ninth and Sixth for the states (the Ninth for the tribes too), and the Supreme Court is the venue that matters. Outside America the argument was never seriously in doubt. Any strategy built on this product needs to be built for at least two of the three outcomes above, because nobody yet knows which one arrives.