Kalshi’s Federal Exchange Hits a State Line

Kalshi runs a federally regulated exchange. In September 2026, the Sixth Circuit nonetheless held that its sports-event contracts were not swaps subject exclusively to CFTC oversight, leaving states able to regulate them under gambling laws. The trading screen could reach a customer whom the exchange could not lawfully serve with that contract.

In 2024, Robinhood offered presidential contracts to eligible US users through ForecastEx. The arrangement made prediction markets look like a distribution problem: write an intelligible question, place it inside a familiar brokerage app, and let more people trade their views. But a contract is not a published forecast. Someone must have authority to offer it, find a counterparty, hold the risk and decide what earns the payout.

Prediction markets began by making probabilities easier to trade; as brokerages, media surfaces and AI agents widen the entrances, the scarce asset is becoming the venue’s capacity to list a permissible contract, concentrate liquidity, clear positions, police participation and establish an accepted settlement. The available integrations bring orders to market operators. They do not show distributors taking over those obligations.

A federal exchange still has to cross a state line

The Sixth Circuit ruling concerned Kalshi’s sports contracts, not a judicial classification of every prediction market. It also did not end the legal fight: reporting around the decision described two federal appeals courts favoring state authority and one favoring prediction-market companies. A national audience therefore cannot be treated as a single reachable market for every product.

A Washington judge had already ordered Kalshi to stop offering most prediction contracts in the state after finding likely violations of state gambling-related laws. New York’s attorney general sued Polymarket US in September, alleging that it operated an illegal gambling business; that allegation remains a claim in litigation, not a finding. These are different proceedings with different scopes, but each forces a venue to answer a jurisdictional question before a brokerage or agent can submit an order. The broader regulatory turn in prediction markets may constrain an incumbent as readily as it protects one.

A separate September report said the Commerce Department had ordered Kalshi to remove an AI-compute futures product and that Kalshi complied. The alleged Commerce Department order appears in four coverage entries, all classified as rumored. The report raises a product-level question about what an exchange may list; it does not establish that a federal veto occurred.

Institutional hedgers need a book, not just an audience

Kalshi says its institutional trading volume has grown 800% since November 2025 as businesses began using its contracts to hedge risks. Kalshi did not supply a starting volume in the cited report, so that growth rate alone cannot establish the size of its institutional business or the depth of a particular market. It does explain why an institution would care about the orders waiting on the other side of its trade, rather than the number of people who saw a quote.

Kalshi Klear, the company’s internal clearing house, has filed for CFTC approval to add margin trading as Kalshi seeks institutional liquidity. Approval would change how participants fund and manage positions; a new entry point in a consumer app would not do that work. Pew’s study of 11,989 Polymarket accounts found that 61% placed fewer than 100 trades over six weeks. That study does not measure order-book depth, but it cautions against treating broad participation as continuously available counterparties. The economic contest between venues is over liquidity that traders can actually use.

Kalshi appeared in enterprise-framed coverage in 25.0% of its 2024–2026 articles, versus 0.0% in an earlier comparison period of just three articles. Those shares measure coverage, not revenue, and the small earlier sample cannot establish when the business changed. Kalshi’s margin filing supplies the more concrete evidence of what institutional participation asks the operator to build.

A weather forecast becomes a trade only when someone can settle it

Kalshi expects $1.1 billion in weather-market trading volume in 2026 and has partnered with Weather Co. to verify settlements for climate and weather contracts. The figure is Kalshi’s forecast, not completed volume. The partnership addresses a different problem from finding traders: a weather outcome must be matched to the observation and rule specified in the contract. A useful meteorological forecast cannot, by itself, decide which reading governs a payout.

That distinction gives contract wording material consequences. Prediction-market disputes have turned on linguistic technicalities, with a single word affecting whether a bet pays. Weather Co. can help verify an observation; Kalshi must still stand behind the market terms under which that observation counts. The venue’s rules give a forecast the narrower, enforceable meaning of a trade.

Kalshi also has to govern who may profit from a contract. It issued its first lifetime ban to George Santos after the CFTC alleged that he made $17,000 betting on his own State of the Union appearance. An accurate price cannot resolve an allegation of improper participation. Kalshi had to make a participation decision, separate from deciding the event’s outcome.

Robinhood can gain leverage without becoming the referee

Robinhood added Crypto.com’s yes-or-no contracts and took minority stakes in Crypto.com and prediction-market company OG. That is stronger than merely displaying another company’s odds. Robinhood owns a customer relationship and has bought an interest in suppliers; substantial distribution could give it leverage over which contracts customers see and on what terms.

Retail traders have also begun training AI agents to trade for them, while Polymarket and Bybit have introduced agent-friendly interfaces. Agents that compare venues could make the user-facing layer more substitutable, and River Markets is already giving professional traders one place to trade across multiple prediction markets. Those developments put pressure on venues to win orders. They do not make two differently worded contracts identical or move an unsettled position’s obligations to the software that routed it. Neither Polymarket nor Bybit has demonstrated broadly deployed, standardized event-contract APIs that would make venue access interchangeable.

Perpetual trading gives the venue a different clock to keep

Kalshi has announced plans for CFTC-regulated perpetual futures, while reporting says it intends to seek approval for single-stock perpetuals; Coinbase has filed similar paperwork. Hyperliquid shows the commercial attraction of continuous trading: amid the Iran war, cumulative volume in its perpetual oil futures rose from about $339 million on February 28 to roughly $7.3 billion on March 12. Kalshi cannot determine from Hyperliquid’s volume figures whether perpetuals will earn it more than event contracts. Those volumes do show why an exchange built around resolving discrete questions would pursue a product that can keep trading.

Continuous contracts replace a single event’s resolution with ongoing demands on contract design, price references, risk management and uptime. CME Group and Silicon Data have launched computing-capacity futures tied to daily GPU rental-rate benchmarks—an example of turning an economic exposure into a specified, tradeable reference. Benchmark design is only part of the job. In November 2025, a cooling failure at a CyrusOne data center halted CME futures trading, including S&P 500 and Nasdaq futures. The trade fit on a screen; its continuity depended on a cooling system.

CME’s outage exposed the machinery behind a quote. Kalshi’s September ruling exposed a different boundary: a sports contract may appear on a national screen where state law bars the exchange from offering it.