Kalshi Says It’s Not Like Predatory Sportsbooks. NY’s AG Wants Execs to Prove It.

Prediction markets like Kalshi and Polymarket often pitch themselves as a cleaner alternative to traditional sportsbooks, matching buyers and sellers in a peer-to-peer marketplace rather than acting as the casino house that profits when customers lose. But New York Attorney General Letitia James is now pressing the company to show that distinction is more than marketing language. In a verified petition filed in New York County Supreme Court, James seeks to enforce a September 25th subpoena issued as part of a civil investigation into potential fraud and deceptive advertising in the prediction market industry. According to the filing, Kalshi has produced zero documents to date. The investigation examines marketing materials, payments to influencers on platforms like Discord and TikTok, and the role of Kalshi’s own affiliated trading arm and third-party market makers on the platform. It sits separate from the state’s earlier lawsuit alleging Kalshi operates an illegal unlicensed gambling operation that allows users under 21, the minimum age for mobile sports betting in New York. The state also sued Polymarket in September for operating as an unlicensed gambling operation. Tarek Mansour has for days now said that what sets him apart from casinos is that he doesn't win if you lose But the NY AG now says that actually the other side of your parlay bet is literally Kalshi Trading! https://t.co/uChs1Sm1Rb — Saagar Enjeti (@esaagar) November 3, 2026 The petition raises concerns regarding Kalshi’s “combo” products, which effectively act as the platform’s version of the multi-leg parlays long offered and heavily promoted by sportsbooks. According to the filing, 100% of those combo bets have either Kalshi Trading, which is the company’s own affiliated market-making entity, or one of Kalshi’s approved professional market makers on the other side of the trade, not another ordinary retail user. An analysis by InGame found that combo takers have lost more than $1 billion since the feature launched in September 2025, with roughly $608.9 million in prediction losses combined with another $400 million paid in fees. Market makers earned approximately $519.8 million after fees on those trades. Kalshi co-founder and CEO Tarek Mansour has repeatedly drawn a clear distinction between his platform and sportsbooks. In a December 2025 appearance on CNBC’s Mad Money, he said, “On Kalshi there is no house, anybody can win.” In an April 2026 interview with Axios, he stated that sportsbooks are “designed for customers to lose” and that when customers win, “they block that customer because those winnings are coming from the business model.” He has argued that no such limitation on winning bettors occurs on Kalshi because the platform is a neutral exchange that simply takes a fee regardless of who wins. Are Market Makers Just Sportsbooks By Another Name? The structure on combos raises questions about how neutral that exchange remains. Combos are filled through a request-for-quote system that allows approved market makers, including institutional firms, to provide the opposing side almost instantly. Ordinary users generally cannot take the other side of someone else’s combo. The arrangement echoes a pattern familiar from the crypto industry, where projects advertise decentralization while concentrating key functions with a handful of professional or affiliated actors. For example, stablecoin issuers and various crypto projects acted more like traditional banks after the recent $6.3 billion Binance hack, freezing roughly $3 billion worth of assets. Legendary NBA bettor Haralabos Voulgaris, who also owns Spanish football club CD Castellón, highlighted a further concern on X. Citing the InGame figures, he noted that market makers appear able to recognize repeat users through persistent account identifiers and can choose whether to quote them at all. Kalshi has said it never limits customers for winning, but Voulgaris questioned the practical value of that policy if professional firms can simply stop providing liquidity once they identify a skilled or unprofitable-to-serve customer, while continuing to offer prices to those who consistently lose. Voulgaris added: “As someone who ‘gambled’ for 20+ years, I can assure you it’s gambling. And a particularly insidious form when you obscure how predatory it is by rebranding parlays as ‘combos’ and betting as ‘trading.’ These sites can and should do better.” He later pushed back against the idea that accounts can simply anonymize their IDs to prevent this issue, posting data to illustrate that liquidity dries up for users whose IDs are hidden. DraftKings co-founder Matt Kalish also previously noted that Kalshi shares user identifiers in its market-maker API in a way that allows identification of winning traders. Prediction Markets’ Endless Controversies The New York filing is only the latest front in a broader fight between states and the Commodity Futures Trading Commission over who regulates the sports contracts on these platforms. The CFTC intervened in New York’s aforementioned separate lawsuit on this point of contention, ordering the company to keep operating in the state. Utah recently took the dispute to another level with criminal charges against Kalshi and its co-founders. The prediction market platforms maintain that their event contracts are federally regulated swaps under the CFTC’s exclusive jurisdiction, not state-regulated bets. There are also recurring questions around insider trading. Some of the more notable cases here involve former Representative George Santos allegedly betting on Kalshi using nonpublic information about his own plans to attend the State of the Union, a Google engineer charged with using confidential search data to profit more than $1 million on Polymarket, political campaign staffers placing bets of their own, and an Army soldier charged in connection with trades on Polymarket ahead of the capture of Venezuelan leader Nicolás Maduro. Sports have driven much of the recent volume growth that has put prediction markets in regulators’ sights, but election-related activity, which is what originally brought these platforms to prominence, has also expanded sharply. According to Artemis, Election Day trading volume has grown from roughly $10 million to about $4 billion since 2020. That said, a report from earlier this year found that sports contracts accounted for about 76–80% of Kalshi’s trading volume. The real winner last night: prediction markets-While not all volume is politics related, growth from $10m -> ~$4.0b since 2020 Over $10b for 2028 presidential race? pic.twitter.com/4q4eK88XS6 — Artemis (@artemis) November 4, 2026 Supporters of prediction markets frequently argue that the prices they produce offer more accurate, continuously updated probabilities than traditional forecasts. Indeed, a Federal Reserve report from February previously found that Kalshi markets provided comparable or better predictions than surveys and other derivatives for certain economic releases, including CPI data and federal funds rate decisions. According to Kalshi, traders on their platform correctly favored the eventual winner in 479 of 487 midterm races on Tuesday, a 98% hit rate the company said beat both traditional polls and forecasting models. But prediction markets also missed badly on Brazil’s recent presidential runoff, pricing Flávio Bolsonaro near 80% on Polymarket and Kalshi before Luiz Inácio Lula da Silva won.