The Circuit Split Is the Headline. The District Courts Are the Story.

Editor's note: TrueEdge builds odds tools and earns affiliate commissions from licensed sportsbooks, which compete directly with the exchanges this piece is about. Our editorial position has not changed and is against our commercial interest: criminal exposure is the wrong instrument to point at a federally registered exchange. What follows is an argument about where this fight is actually being decided. Weigh the conflict.

Judge Stephen H. Locher did not need the Commodity Exchange Act to decide the Iowa case. He needed Kalshi's App Store copy.

In denying the exchange's motion for a preliminary injunction, the Southern District of Iowa noted that Kalshi had marketed its product as "the first app for legal sports betting in all 50 states," and treated that as evidence of what the product is. Then he wrote the sentence that should worry Kalshi more than anything the Ninth Circuit said: "If Congress intended to pre-empt state gambling laws when it enacted and amended the Commodity Exchange Act, it should have been clear about that pre-emption. The text of the Commodity Exchange Act does not meet the mark."

That is not a close call dressed up as a holding. That is a judge saying the statute does not contain the thing Kalshi says is in it.

Two losses in one day, and neither was the one that got covered

Tuesday, September 8 was the worst day Kalshi has had in court, and almost nobody wrote it that way, because the same week produced a cert petition and an en banc petition and those are the filings that generate headlines.

The Tenth Circuit denied Kalshi's emergency motion for an injunction pending appeal in the Utah case in a two-paragraph order. Not a narrow denial on one factor — all four. The panel found Kalshi had not shown likelihood of success, had not shown irreparable harm, had not shown the balance of harms favoured it, and had not shown the public interest did. "Kalshi has not shown these factors weigh in its favor. Accordingly, we deny its motion." That was the whole analysis. An appellate court that thought the question was hard would have written more.

Utah matters because of what the statute says. Offering online gambling to any individual in the state is a third-degree felony there — not a licensing violation, not a civil penalty. Kalshi filed suit in February, after Governor Spencer Cox and Attorney General Derek Brown made clear the state viewed sports event contracts as illegal gambling. Judge Robert J. Shelby denied the preliminary injunction at the district level in August. The Tenth Circuit has now declined to freeze that while the appeal runs. Utah's route from here looks like Connecticut's: a civil enforcement action, filed after the federal courts declined to protect the exchange. Connecticut brought one in August and a state court declined to issue an immediate temporary restraining order — which is to say the state's suit survived its first contact with a judge, and Kalshi's shield did not.

Iowa came down the same week, and we could not get a docket number for it from any outlet we could reach, which is worth saying out loud. What we can report is the reasoning, and the reasoning is the part that travels. Locher concluded that sports event contracts are not "swaps" within the meaning of the Act, and that the CFTC's exclusive jurisdiction therefore does not reach them. Kalshi had sued Attorney General Brenna Bird and the Iowa Racing and Gaming Commission after a March 4 meeting at which, by the company's own account, Bird made clear she considered the contracts illegal under Iowa law.

Then Illinois picked up the Iowa order on September 9 and filed it as supplemental authority against Kalshi's pending motions there. That is how a district-court loss becomes a national one. It takes about twenty-four hours.

The convergence is the signal, not the split

Here is the argument. The circuit split is real and it is being covered accurately, and it is also, for the next eighteen months, mostly theatre.

The Third Circuit ruled for Kalshi against New Jersey in April. The Ninth Circuit ruled for Nevada on August 28, 3-0, with the panel finding that the contracts "were not 'swaps' because they were sports bets." Kalshi petitioned for rehearing en banc on September 9, arguing the decision creates a split on an exceptionally important question of federal preemption and exposes it to civil and criminal liability while it is "abiding by federal law and the mandate of its exclusive federal regulator." Robinhood and Crypto.com filed for certiorari on September 10. New Jersey's own regulators have a petition in. The Supreme Court will decide whether to take one of them, and if it does, FanDuel's chief executive told a Bloomberg audience this week he expects a decision in late 2027 or early 2028.

Eighteen months is a very long time for a company whose product is only worth what it is worth if it can be offered everywhere at once.

And in that window, the thing that determines Kalshi's operating footprint is not the appellate map. It is what trial judges do with preliminary-injunction motions, one state at a time. Look at what they are doing. Locher in Iowa: not swaps, no field preemption, and your marketing says you know it. Nelson's Ninth Circuit panel: not swaps. Shelby in Utah, upheld by the Tenth Circuit's refusal to intervene. Judge Donald W. Molloy in Montana on August 27, denying the injunction on the ground that "Kalshi's repeated agreement to stay this matter for months at a time fatally undermines its ability to make such a showing" of irreparable harm.

Twenty states are in active litigation with prediction-market operators. Forty-three signed on to Ohio Attorney General Andy Wilson's July letter rejecting the CFTC's claim of exclusive jurisdiction. The exchanges are not facing a split. They are facing a consensus with one appellate exception.

The strongest case against this reading

A serious version of the other side goes like this, and it has three parts, two of which are correct.

First: preliminary-injunction denials are not merits rulings. A judge finding a movant unlikely to succeed is making a forecast on a compressed record, and forecasts get revised. That is true, and we should not pretend a PI denial is a final judgment.

Second: counting Molloy's Montana ruling in a preemption losing streak is sloppy, and we will not do it. Molloy denied on irreparable harm, and specifically on the ground that Kalshi had repeatedly agreed to stay its own case — a laches-shaped ruling about litigation conduct, not a holding about the Commodity Exchange Act. He denied it without prejudice. It belongs in a story about Kalshi's strategy, not a story about the statute.

Third, and this is the part we think is wrong: that the Third Circuit plus a likely cert grant makes all of this temporary. It might. But the preliminary posture cuts the opposite way from how the exchanges are using it. A preliminary injunction exists to preserve the status quo for a party that will otherwise suffer harm it cannot be paid back for. Kalshi has now been told by a district judge in Utah, an appellate panel in the Tenth Circuit, a district judge in Iowa and a district judge in Montana that it has not made that showing. When four courts in five weeks decline to preserve your ability to operate, "this is only preliminary" stops being a defence and starts being a description of how long you will be switched off.

We do not know whether the Supreme Court will grant cert, and anyone telling you they do is guessing. We also do not know whether the Iowa and Utah reasoning survives full merits briefing. What we can say is that the "not a swap" conclusion was reached independently by a Ninth Circuit panel and by a district judge in Des Moines working from the same statutory text, and independent convergence on a reading is the best predictor available of where the merits land.

What would change our mind

An en banc grant at the Ninth Circuit. Those are rare — one practitioner's count puts it at 11 of 730 recent petitions — and a grant would mean at least one active judge thinks the panel got the statute wrong, which is a materially different signal than a cert petition nobody has ruled on.

The other thing that would change it is a district court, anywhere, denying a state's enforcement position on the merits after full briefing. That has not happened since April.

The position we have argued since this beat opened is unchanged, and it is not the position our balance sheet prefers: a state that believes a CFTC-registered exchange is breaking its law should say so in a civil action, not reach for a felony statute. Utah's third-degree felony exposure is the wrong instrument regardless of who wins the preemption question. But arguing that the criminal lever is wrong is a different thing from arguing that the preemption case is strong, and this week made the second argument considerably harder to make with a straight face.