Is Kalshi Gambling? Prediction Markets, Sports Bets, and the Fight Over Who Regulates Them

The same bet, in a different suit

You self-excluded from every sportsbook. You handed over the passwords. And then a friend mentions Kalshi, or you open a finance app and there's a tab that lets you buy “yes” on tonight's game. It looks nothing like a betting slip. It's priced like a stock. Whether the law treats it as a federally regulated contract or a state-regulated wager is now the subject of conflicting federal appeals-court decisions. That gap — between the product label, the user experience, and the legal regime — is the whole story.

Kalshi and Polymarket are the two biggest “prediction markets.” What they sell is an event contract: a yes/no question about whether something will happen — a game, an election, an economic number — priced between about a penny and 99 cents, that pays a dollar if you're right and nothing if you're wrong. Buy “yes” at 60 cents and the payoff is arithmetically similar to a bet at odds. Kalshi and the CFTC argue that contracts listed on a federally registered exchange belong under commodities law. State regulators argue that sports versions remain wagers under their gambling laws.

The financial-contract structure is the product, but its legal consequences are contested. That matters to a person using self-exclusion or other safeguards, because protections tied to licensed gambling do not automatically follow every event-contract platform.

For the person betting, the house math is the same

Start with scale, because it explains why this stopped being a curiosity. Combined monthly trading volume on Kalshi and Polymarket went from under $5 billion in September 2025 to about $24 billion by April 2026 — approaching the size of the entire US legal sportsbook market. On Kalshi, the large majority of that volume — roughly 80% since mid-2024, and higher in some recent months — is sports.

Now the part the marketing skips. When researchers at Yale studied 1.72 million Polymarket accounts, they found the accuracy these markets are praised for is produced by a tiny minority of skilled traders. The rest, the Yale team wrote, “does not produce accuracy; rather, it funds it.”

That is the loss curve of a house game, not a savings account: a few winners, funded by everyone else. Prediction-market prices really can beat polls close to an event — that's true, and worth conceding. But “the price is often right” is a different claim from “you will make money,” and for the typical retail user the second one doesn't hold.

“

The label says market. The math says house.

Who regulates it — and why nobody agrees

Here's the honest legal picture, because it matters. Kalshi and the CFTC argue that sports event contracts traded on a CFTC-registered designated contract market are “swaps” under federal commodities law and therefore sit within the CFTC's exclusive jurisdiction. State regulators say the same products are wagers subject to their gambling laws. On April 6, 2026, the Third Circuit ruled on that conflict. Its 2–1 majority affirmed preliminary relief in New Jersey after finding that Kalshi had a reasonable chance of showing federal preemption for the sports contracts at issue. That was not a final merits judgment or a nationwide rule.

On August 28, the Ninth Circuit reached the opposite preliminary-stage result in Nevada. It affirmed the dissolution of an injunction that had stopped Nevada from enforcing its laws against Kalshi's sports contracts, holding that Kalshi had not shown it was likely to establish federal preemption. The court sent Nevada's challenge to Kalshi's election contracts back to the district court without deciding that question. That decision is also not a final nationwide judgment.

The federal regulator is still pressing its own position. The CFTC has sued states including Arizona, Connecticut, and Illinois, arguing that their enforcement intrudes on federal jurisdiction. It has also proposed a rule that would define terms, apply public-interest factors, and create a 90-day review process for contracts involving gaming and other enumerated activities. The proposal treats some aggregate sports outcomes more favourably than contracts involving injuries, officiating, altercations, or youth sports. As of August 31, 2026, it remains contract-specific, creates no safe harbour, and is not a final rule.

So the answer now depends on what question you are asking. The Third Circuit and Ninth Circuit read the federal-state boundary differently in two preliminary disputes involving different states. The CFTC and state regulators remain on opposing sides. Sports and election contracts can require separate analysis. Anyone telling you one court has made every prediction market legal — or illegal — everywhere is skipping the actual decisions.

When a platform says “this is a regulated market, not a sportsbook,” it is stating one side of a live legal dispute. CFTC oversight and enforcement exist, but the label is not a finding that the product carries the same consumer protections as a licensed sportsbook. Jurisdiction answers who regulates; it does not answer whether the product is a safe fit for a particular person.

This article provides general information, not legal advice. Court and agency proceedings can change after the August 31, 2026 update date.

Insiders, whales, and manufactured odds

The pitch for prediction markets is that the crowd is wise — the price is the probability. That breaks in exactly the ways that hurt the person trusting it.

Start with insiders, because 2026 was the year they began getting caught. In April, federal prosecutors charged a US Army soldier with using classified information about a Venezuela military operation to win roughly $400,000 on Polymarket. A month later the CFTC — alongside Manhattan prosecutors — accused a Google engineer of using nonpublic “Year in Search” data to make about $1.2 million, also on Polymarket. Both are accusations the men are presumed innocent of. Kalshi, for its part, fined a MrBeast video editor $15,000 and suspended him two years for trading on inside information, and says it opened roughly 200 insider-trading investigations in a single year.

Then there's manipulation of the market itself. A 2025 Columbia working paper estimated that about a quarter of Polymarket's historical volume — and nearly half of its sports volume — was “wash trading,” accounts trading with themselves to inflate activity. It's an estimate, not a verdict, and Polymarket disputes it. But the structural point stands: a thin market can be moved by one whale, as a single trader visibly did to the 2024 US election odds.

And the number can be wrong even after the event. Because Polymarket settles outcomes by a token-holder vote, results can diverge from reality: in March 2025 a market on a Ukraine “mineral deal” resolved “yes” — and paid out — even though no deal existed, after what Polymarket itself called an “unprecedented” governance attack. It refused refunds.

None of this makes Kalshi or Polymarket a scam; both are real companies. Polymarket has the regulatory record to match — a 2022 CFTC order fined it $1.4 million and pushed it to block US users, the FBI searched its CEO's home in late 2024, and those federal investigations closed in July 2025 with no charges. The point is narrower and worse for a bettor: the “probability” on the screen is only as honest as the thinnest, most anonymous corner of the market — and you're not the one who can see into it.

The guardrails are different, not absent

Here's where the legal label stops being an abstraction. State self-exclusion programs and gambling-specific controls are generally built around the operators and products covered by that jurisdiction. They do not necessarily extend to a federally regulated event-contract platform. Private and user-declared tools can still target named services: Gamban, a gambling-blocking vendor, added event contracts to its blocklist in late 2025 and said the quiet part plainly: in its view, calling a bet a “trade” can leave self-exclusion, deposit-limit, and age-verification rules dependent on a different regulatory regime.

There is also a voluntary private self-exclusion program for prediction markets. As of August 31, 2026, SelfExclude lists Kalshi and Novig as integrated and describes other platform integrations as in progress. That program exists, but it is voluntary, private, U.S.-only, and limited to participating platforms and its stated terms. The National Council on Problem Gambling argues that event-contract platforms should carry stronger consumer protections and describes the activity as “functionally gambling, regardless of how it is legally defined.” That is an advocacy position, not a court holding, and NCPG says the longitudinal evidence specific to prediction markets is still emerging.

And it's going mainstream. Robinhood built a large event-contracts business; when it launched Super Bowl contracts in February 2025, the CFTC formally asked it to stop, and it did — a day later — before returning with March Madness contracts weeks after. In Canada, Wealthsimple Predict brought the same machinery to a mainstream investing brand: a standalone app, regulated by CIRO and powered by Kalshi, offering binary yes/no contracts — though Canadian regulators drew a harder line, limiting it to economic, financial, and climate events and barring sports and politics. The bet button is moving off fringe crypto sites and into the app that also holds your savings.

If this is your trigger

None of this means the markets are lawless. Federal market rules, platform controls, enforcement actions, and private self-exclusion all exist. The narrower point is that coverage differs: a state gambling registry, a platform control, and a private cross-platform list are not the same thing. If prediction markets are your trigger, waiting for the courts to settle the label does not create a personal boundary.

One voluntary layer is a boundary you declare yourself — naming the platforms you never want to reach in GuardianBlock Custom Blocks. A protected adult can add a personally declared no-go domain, while later removal or pause follows keyholder approval. That is one friction and accountability layer on an enrolled device, not proof that every route or platform is covered, and it complements rather than replaces account closure, applicable self-exclusion, financial controls, or professional help.

For the age and consumer-protection question, read our research on Kalshi, young adults, and the NFL. For the practical website boundary, read what it means to block Kalshi and Polymarket on Windows—and what a domain rule cannot cover.

If you're struggling, help exists and it's free. In the US, the National Problem Gambling Helpline (1-800-MY-RESET) is confidential and open 24/7; in Canada, the Responsible Gambling Council lists a problem-gambling help line for every province. You don't have to have lost everything to use them.