- ▸ Core features like shared liquidity, market makers, and order books are restricted or outright prohibited under state gambling regulations.
- ▸ GGR-based taxes and fragmented licensing create costs that exceed exchange-level margins.
- ▸ Operators are moving toward CFTC-regulated models where exchange infrastructure can actually function, out of necessity.
The courts are split on one of the most consequential questions in prediction markets: whether states have the right to regulate exchanges like Kalshi that offer sports event contracts under a federal CFTC license. What’s getting lost is the more fundamental issue that state gambling frameworks weren’t built for exchanges at all. Alex Kane, CEO of Sporttrade and the only founder to have built a state-regulated sports prediction market exchange at scale, has quantified it: To operate across all 38 states with legal sports betting, Kane says, would cost more than $100 million upfront, $50 million or more annually in licensing and market access costs, and the addition of 100-200 full-time staff “to maintain compliance with 38 different technical/information security audits, 38 different testing certifications, 38 different financial audits, and 38 different sets of rules and reporting requirements.”
Those costs reflect something more fundamental, Kane says: “States are unable to regulate exchanges. They never contemplated that model.”
Peter Hammon, an advisor and attorney in the online gaming/prediction markets industry, has helped build the regulatory architecture for two state-regulated sports betting exchanges, Smarkets and Sporttrade. His advice to anyone considering the same path: “I would not advise pursuing that business model under state regulation anymore,” he told DeFi Rate.
Kane and Hammon aren’t describing a regulatory headache, but a structural incompatibility between state gambling law and the exchange model, compounded by equally problematic tax frameworks. As more than a dozen states move to restrict or tax federally licensed exchanges, the Third Circuit ruled April 6 that New Jersey cannot enforce its gambling laws against Kalshi, the strongest judicial endorsement yet of a single federal framework. The Ninth Circuit, however, allowed Nevada to do the opposite. The split could take years to resolve.
In the meantime, the practical constraints are already clear. As Hammon recently put it, there has been “not nearly enough discussion” about the practical impossibility of running a state-regulated prediction market in the US. That starts with how exchanges actually work.
What makes an exchange different
A traditional sportsbook sets prices and profits when you lose. An exchange, the model used by Kalshi and Sporttrade, is a marketplace where participants post bids and offers against each other. The platform takes a small commission on matched trades and is indifferent to the outcome. Liquidity depends on market makers: professional counterparties who continuously quote prices on both sides. With no market makers, there’s no liquidity. And with no liquidity, there’s no exchange. But state gaming regulations expressly prohibit key exchange infrastructure.
State gambling law was written to regulate bookmakers, not exchanges. When Kane built Sporttrade under state rules, the result was five separate exchanges across New Jersey, Colorado, Arizona, Iowa, and Virginia — each with its own order book, matching engine, and licensed liquidity providers, each legally unable to match a customer in one state against a customer in another.
When federally-licensed competitors launched nationally with unified liquidity pools, Kane wrote to the CFTC in April 2025 citing “irreparable harm” and pivoted toward federal designated contract market status. “We decided to view the disruption as a positive for customers, and thus a positive for our vision,” he told DeFi Rate.
The pivot is becoming an industry pattern. Kane doesn’t fault the regulators, but he points to a structural limitation: “I respect the state-regulated system for what it is, and I respect the regulators personally. I think the regulated products work (albeit not for everyone), and they create tax revenue,” says Kane. “They never contemplated that model, as evidenced by the lack of even defining the core parts of an exchange ecosystem (market makers, matching engine, orderbook, brokers, etc).”
Hammon sees the same shift playing out across the industry: “None of these costs are currently imposed on prediction markets — which is why you see so many sportsbooks and exchanges making the leap from state regulation to CFTC regulation.”
The Phantom tax problem: Being taxed on money you never made
How states choose to tax prediction market exchanges, if they’re permitted to do so at all, is an entirely untested question. And the regimes now advancing through state legislatures suggest that some of them would make operating untenable. The structural reason comes down to how exchanges earn revenue versus how states have traditionally measured it in the gaming industry.
In the UK, the Gambling Commission taxes betting exchanges on their commission revenue, the maker/taker fees the platform actually earns. In the US, state gambling commissions tax exchanges on total customer losses (Gross Gaming Revenue). On a betting exchange, most of that money flows directly to the market maker on the other side of the trade. The platform only keeps the commission.
“It’s an important distinction because most of the revenue from customer losses accrues to market makers, so exchanges wind up getting taxed on phantom revenue,” Hammon explained.
Iowa’s proposed SF 2470, which passed the Senate 45-1 and prompted Kalshi to sue the state before it even cleared, is the starkest example of what an unworkable tax regime looks like. It includes a 20% GGR-adjacent levy plus a 20% excise on every single event contract. Kentucky’s HB 757, currently on the governor’s desk, takes a more principled approach by taxing transaction fees rather than customer losses, which is structurally closer to the UK model. It’s a meaningful improvement on Problem One. But with eleven states now considering their own frameworks, it’s possible that other states follow Iowa’s approach.
Hammon says the Iowa tax provision “has no serious chances of ever being enforced” adding that, “The CFTC would immediately sue if Iowa actually tried to collect fees from prediction markets regulated by the CFTC. As a practical matter, no prediction market is going to hand the Iowa Racing and Gaming Commission a check for $20 million.”
The $20 million figure also illustrates the structural absurdity of applying state licensing economics to a federally-licensed industry: “Iowa’s $20 million upfront license fee is a Don Quixote proposal,” Hammon explains. “If expanded to other states at the same rate, that would equal $2 billion dollars per prediction market operator in Year 1. Iowa is 1% of the U.S. population.”
Kane calls state GGR-based proposals “completely unworkable” based on underlying commission economics: “The long-term commission per contract traded on prediction markets is probably 0.1% to 0.3%.” Taxed on GGR rather than actual commission revenue, those margins face a tax bill calculated at multiples of what the platform earned from each trade.
Why institutional liquidity can’t enter state-regulated exchanges
The UK allows corporate entities to market make on betting exchanges without a gambling license. However, US states require market makers to hold the same licenses as casino equipment vendors, which require six-to-seven figures per state, years to obtain, with exhaustive financial disclosure of the C-suite and investors.
Hammon details the workaround that emerged: “States don’t allow for entities to bet, so at Sporttrade, we’ve had to work with the states to create a concept that market makers are gaming ‘suppliers’ or ‘vendors’ to Sporttrade rather than active participants. It’s a creative workaround, but it exposes market makers to an extensive licensing investigation that costs six figures and takes months or years to complete.”
“As you can imagine, that has scared off most wall street firms that otherwise would love to participate,” says Hammon.
Kane identifies two foundational requirements that state law prohibits outright: end users trading electronically, and entities participating as customers. On federally-regulated exchanges, institutional market makers, which are organizations, not individuals, are the primary source of liquidity. But state law doesn’t recognize them as eligible participants.
Kane noted that while Sporttrade does have two licensed service providers that supply liquidity to the exchange, “such a licensing process is so onerous that it has eliminated the 99% of market makers that participate on venues federally, but can’t on Sporttrade.”
“Many of such participants have inquired about joining Sporttrade, but went cold when they learned about the licensing process,” Kane told DeFi Rate.
Every state is its own island
The Wire Act bars sports wagering transactions from crossing state lines, forcing each state’s order book to operate in isolation.
“Combined with state laws that require sports betting servers to exist in each state, these rules mean a company like Sporttrade is obligated to set up a different server in each state,” Hammon told DeFi Rate. “Each server must have its own matching engine, the tech that matches bettors against other bettors or against market makers.”
Hammon explained the dire effect of Wire Act limitations on basic functions of the exchange model: “From a tech standpoint, this means…artificially capping liquidity, and losing the pricing, liquidity, and other benefits that draw customers to a betting exchange/prediction market in the first place.”
Then there’s the issue of having to comply with 38 different regulatory regimes for licensing, operation, reporting and more, a requirement that Kane calls “completely unworkable for any startup.”
“It’s why there are now zero ‘startups’ that operate anywhere in the regulated market, outside of Sporttrade.” The implication here is that a system that blocks new entrants doesn’t just limit competition, but also constrains innovation.
Why a license still wouldn’t be enough
Hammon’s three structural problems are about misalignment between how exchanges work and what state sports betting regulations allow. Anthony Busacca, CFO of Sporttrade, makes a more fundamental point, namely that the regulations don’t just make the exchange model difficult, but prohibit it outright.
“States don’t allow entities to bet, don’t allow algorithmic/bot trading, most don’t allow players to offer their own price, State operators are allowed to limit/ban winners, the list goes on. So it’s not as easy as saying Kalshi can just get a license in each State, as the regulations/statutes don’t allow for the model. So getting licensed would still mean they couldn’t operate a true exchange model at the state level.”
Those prohibitions target key pillars of trading exchanges. We already covered the issue of “no entity betting,” which undercuts the market making engine. No algorithmic trading means institutional market makers can’t operate their core infrastructure, and without continuous high-frequency quoting, spreads widen and liquidity collapses. No limit orders, “most don’t allow players to offer their own price,” could be the most detrimental prohibition of all as limit orders are the foundation of the traditional order book and an integral part of price discovery. And the ability to ban winning players, standard practice for state-licensed sportsbooks, directly codifies the bookmaker model into regulation and contradicts the exchange’s core promise of fairness to all participants.
Winning the preemption argument clears state law out of the way and replaces it with a framework that, as Kane discovered, already has the right foundations for exchanges. When Sporttrade began its DCM application, he says, “it became immediately clear that the CFTC already had rules and policies for how participants could connect, how entities can participate, how intermediation would work — all things that are prerequisites to effectively regulating exchanges.”
The CFTC’s pending rulemaking addresses a narrower set of open questions like which event contracts are permissible and what restrictions apply to sports-related contracts specifically rather than the exchange mechanics that federal regulation already covers.
Why this is coming to a head now
The Third Circuit’s April 6 ruling in KalshiEX LLC v. Flaherty, a 2-1 decision, held that Kalshi’s sports-related event contracts are swaps traded on a CFTC-licensed Designated Contract Market, and that the Commodity Exchange Act’s grant of exclusive federal jurisdiction preempts New Jersey’s gambling laws. The court sided with Kalshi’s “narrow framing”: New Jersey wasn’t regulating gambling broadly, it was impermissibly regulating trading on a federally designated market.
It is the strongest judicial signal yet that prediction markets belong under a single federal framework. It is also not the final word. New Jersey can seek en banc rehearing. More critically, the Ninth Circuit went the other way on Nevada last month, clearing that state to enforce its gambling laws against prediction markets. The circuit split is live. The CFTC sued Arizona, Connecticut, and Illinois on April 2 in a parallel effort to lock in federal preemption. A wave of House and Senate proposals with some aimed at banning sports event contracts on CFTC-regulated platforms, adds a congressional dimension that the courts alone can’t resolve.
Hammon thinks it will be 1-2 years before the issue makes it to the Supreme Court, and that “Status quo holds until then, minus a few states that may get picked off like Nevada.”
While he thinks Congressional action is still unlikely, he also says “Kalshi and Polymarket PR campaigns are so aggressive that it’s actually creating a risk there.”
What comes next
The structural problems Hammon and Kane describe don’t wait for the courts to settle the preemption question. If states win, the exchange model faces a statutory landscape that prohibits its core mechanics entirely: the phantom tax regimes, the market-maker restrictions, the Wire Act fragmentation, and the outright statutory prohibitions that Busacca identified all apply simultaneously. The model doesn’t just become difficult but impossible, and even potentially illegal.
If states lose and preemption holds, the picture changes significantly. The CFTC framework already handles the exchange infrastructure, covering entity participation, algo trading, order books, and intermediation. The remaining open question is the upcoming rulemaking on clear contract rules and restrictions. Kentucky’s transaction-fee approach points toward a workable tax model if states retain any taxation rights at all, but Iowa’s approach points in the opposite direction. The UK shows what happens when you get the structure right and then tax it to death.
For Hammon, who has now navigated both paths, the direction is clear:
“I’ve now helped launch two different state-regulated sports betting exchanges — Smarkets and Sporttrade. I would not advise pursuing that business model under state regulation anymore. If you have the tech ready to go, do whatever you can to acquire a CFTC DCM as soon as possible, or look to partner with an existing DCM that may not have sports wagering exchange technology in place.”
In practice, the industry isn’t waiting for the courts. It’s already choosing the model that works.
