The New York Stock Exchange’s parent just completed a $2 billion bet on Polymarket. Kalshi raised a billion at $22 billion while generating fee revenue most exchanges would envy. Seven bills in Congress want the whole category banned. Wall Street is not gambling on prediction markets; it is buying the probability layer of the financial system, and the difference explains everything.
Summary
- Intercontinental Exchange, parent of the NYSE, completed a $2 billion commitment to Polymarket in March, $1 billion in October plus a fresh $600 million, with the platform now discussing new funding near a $15 billion valuation.
- Kalshi raised more than $1 billion this spring at a $22 billion valuation, roughly doubling in months, on volumes that reached $31.5 billion in June against Polymarket’s $10.8 billion, with fee revenue estimates running from $850 million to $1.5 billion annualized.
- The tell is the deal structure: ICE bought global distribution rights to Polymarket’s event data and launched institutional probability feeds within months, chairman Jeffrey Sprecher framing the stake as a new layer of financial intelligence, not a venture flyer.
- The consolidation is visible everywhere: the rival CEOs jointly backed a $35 million VC fund for the sector, Kalshi struck institutional distribution through Tradeweb, Robinhood’s event contracts out-earned its crypto business, and banks project the industry toward $10 billion in annual revenue by 2030.
- All of it is happening against maximal legal hostility: at least seven bills targeting the category in 2026, a bipartisan act to ban sports contracts outright, and the 50-state jurisdictional war this publication has mapped, a contradiction the valuations are pricing as temporary.
Wall Street has a reliable tell: watch what the exchanges buy. Exchanges are the market’s landlords; they monetize activity without taking its risks, and when an exchange operator writes a ten-figure check, it has concluded that a new kind of activity is durable enough to tax.
In March, Intercontinental Exchange, the $80-billion-class operator of the New York Stock Exchange and twelve other regulated venues, completed exactly that judgment: a $600 million investment closing out a $2 billion total commitment to Polymarket, the crypto-native prediction market, at valuations that climbed from $9 billion toward the $15 billion its next round now targets.
Weeks earlier, Kalshi, Polymarket’s regulated arch-rival, raised more than a billion dollars at a $22 billion valuation, double its winter mark, on revenue that estimates place between $850 million and $1.5 billion a year.
The two firms’ founders then jointly seeded a venture fund for their own sector, the corporate equivalent of rival generals founding a military academy, while Robinhood’s earnings quietly revealed that event contracts already out-earn its crypto business. And Congress, watching all of it, introduced at least seven bills to restrict or ban the category.
This piece is the cluster-opener the moment deserves: what the exchanges actually bought, why the valuations disagree with the volumes, and why the industry’s legal peril and its institutional embrace are, strangely, the same story.
What ICE actually bought
The most analyzed deal in the sector is also the most misread, because the commentary priced it as a bet on betting, and the structure says something else.
ICE’s $2 billion did not primarily buy a share of trading fees. It bought, alongside equity, global distribution rights to Polymarket’s event-driven data, and the follow-through arrived within months: Polymarket Signals and Sentiment, launched in February, packages real-time prediction-market pricing into structured feeds for institutional clients, sold through the same ICE data machine that distributes bond pricing and commodity curves to every terminal on earth.
Chairman Jeffrey Sprecher’s framing was explicit and deserves to be taken literally, not as deal-announcement poetry: the investment adds a new layer of financial intelligence. Translated from exchange-operator: markets on events generate a product exchanges have never had, continuously priced probabilities of the world’s discrete outcomes, elections, rate decisions, wars, product launches, and the firm that owns the distribution of those probabilities owns something adjacent to what Bloomberg owns in reference data.
The trading is the factory; the data is the product; and ICE, whose entire modern history is converting exchanges into data companies, ran its signature play on the newest exchange category in existence. The tokenization collaboration attached to the deal, and Polymarket’s acquisition of DeFi infrastructure startup Brahma to harden its on-chain stack, complete the picture of a platform being fitted for institutional plumbing, not fattened for retail.
Read that way, the sector’s other moves stop looking like a bubble and start looking like a supply chain assembling. Kalshi’s deal with Tradeweb points the same direction: prediction pricing distributed into institutional macro workflows, where a Fed-decision market is not a casino but a hedging instrument with cleaner event definition than any rates option.
The $35 million venture fund jointly backed by both CEOs is the infrastructure tell, seeding the tooling layer, the Bloomberg-style terminals, the risk systems that a maturing asset class requires. And Citizens Bank’s projection, a $3 billion industry revenue run rate with a credible path toward $10 billion by 2030, is the sell-side arriving to cover a category it can finally model, because fee-on-volume businesses are the one thing Wall Street knows how to value.
The scoreboard, and why it disagrees with the valuations
Now the numbers, because they contain a genuine puzzle: the market leader by volume is not the platform the world’s biggest exchange bought, and the valuations invert the operating data.
Kalshi leads operationally, and not narrowly. June volume near $31.5 billion against Polymarket’s $10.8 billion, roughly three-to-one; revenue estimates, whether the conservative $850 million fee figure or the $1.5 billion annualized estimates, that would flatter mid-tier traditional exchanges; 994% year-over-year growth into 2025’s $260 million; a CFTC-regulated perch that makes it the only fully domestic, fully legal venue for American event trading; distribution through Robinhood’s hundred-million-account retail machine; and an international expansion running through Brazil.
Its $22 billion valuation, double in months, prices continued category leadership. Polymarket’s operating story is choppier: bigger brand, deeper crypto-native liquidity, the QCEX acquisition purchasing a CFTC-licensed re-entry to the US market it was exiled from, February volume that touched $23 billion in the World Cup’s glow, but June at a third of Kalshi’s tape, revenue still being switched on, taker fees arriving only this spring, and a valuation, $9-10 billion post-ICE, now reaching for $15 billion, that trails its rival’s by a third despite the grander patron.
The resolution of the puzzle is the thesis of this piece. Kalshi is valued as an exchange: volumes, fees, growth, multiply. Polymarket is valued as infrastructure: the ICE relationship prices not its current fee take but its position as the probability layer ICE intends to distribute, the way index providers are valued not on their own revenue but on their irreplaceability in other people’s products.
Two different theories of what a prediction market is, held by two different kinds of capital, produce valuations that disagree with the volume table, and the disagreement is the sector’s live experiment: whether the money is in operating the casino or in owning the odds. The honest answer, visible in ICE’s own history, is that the second usually wins over decades, and the first pays for the war in the meantime.
The paradox: embraced and indicted at once
Everything above happened while the American legal system escalated its hostility, and the contradiction is not a footnote; it is the sector’s defining condition.
The docket, which this publication’s fifty-state-war coverage maps in detail, has only thickened: at least seven bills introduced in 2026 targeting prediction markets, headlined by the bipartisan Prediction Markets Are Gambling Act from Senators Schiff and Curtis, which would ban CFTC-regulated exchanges from listing sports contracts outright; Nevada’s temporary restraining order against Kalshi in March; state gaming regulators across a dozen jurisdictions pressing the position that event contracts on sports are wagers wearing derivative costumes; and a parallel insider-trading discourse, sharpened by suspicious trades around geopolitical events on Polymarket’s offshore book, that has drawn congressional letters.
The CFTC has intervened repeatedly on the federal-preemption side, backing the position that its regulated exchanges answer to it exclusively, the exact architecture, federal license against state police power, whose constitutional endgame our earlier coverage called the sector’s real stakes. Sports is the war’s center of gravity because sports is the revenue’s center of gravity, and a Schiff-Curtis world would amputate the category’s largest limb.
So why do the checks keep clearing? Because institutional capital has concluded the contradiction resolves in one direction, and the reasoning deserves to be stated plainly rather than dismissed as bravado.
First, the federal question is being litigated on ground the industry mostly wins: preemption doctrine has historically favored federally licensed markets, and every court that lets a CFTC venue keep operating strengthens the ratchet.
Second, the political economy shifts with every acquisition: when the NYSE’s parent owns the odds layer, when Robinhood’s earnings depend on event contracts, when Tradeweb distributes the pricing, a ban stops being a consumer-protection bill and becomes an attack on exchange infrastructure, and exchange infrastructure has the best lobbyists money retains.
Third, the data thesis is jurisdiction-proof in a way the trading thesis is not: even a sports-banned, state-harried American prediction market generates probability data on everything else, elections, rates, geopolitics, and the data product ICE bought travels globally regardless of where the trading is licensed.
The valuations, read correctly, are not ignoring the legal war; they are pricing its outcome, a federally consolidated, institutionally owned category whose bumpy state-level litigation is the cost of the moat being built, because the same legal complexity that harasses the incumbents forecloses new entrants. Whether that pricing is right is the decade’s question. That it is the pricing is no longer in doubt, and the November midterms, the category’s largest-ever volume event arriving in the middle of its largest-ever legal fight, will be the first full-scale test of both theses at once.
The third front: retail distribution and the Robinhood wedge
Between Kalshi’s regulated exchange and Polymarket’s crypto-native book runs a third front this publication has tracked from its earnings angle, and the cluster map is incomplete without it: the brokerages, who own the customers everyone else is paying to reach.
Robinhood’s numbers made the case before any thesis did: event-contract revenue of $147 million in a single quarter, up 320%, out-earning the firm’s entire crypto trading business, on 8.8 billion contracts, distribution economics that no standalone platform matches because the marginal customer already has the app, the account, and the balance. The firm’s response to its own data was vertical integration, the Rothera exchange venture, converting Robinhood from Kalshi’s largest distributor into its future competitor, a sequence every platform economist recognizes: distribute, learn the margins, then own the venue.
The World Cup quarter’s industry-wide records, Kalshi clearing $31 billion in a month with retail brokerage flow as a major tributary, previewed the steady state: prediction volume as a standard feature of every retail trading app, the way options became one, with the venues competing for the pipes as much as the punters.
The retail front reframes the legal war’s stakes too. The seven bills and the state actions target venues, but the wedge that actually carried event contracts into American living rooms is the brokerage integration: regulated firms with licenses to protect, distributing CFTC-listed contracts to mass-market accounts, and any legislative ban must therefore run through the brokerage lobby as well as the exchange lobby, a coalition that has historically been the most effective in financial politics.
It also sharpens the insider-trading and consumer-protection critiques, because the retail customer buying an election contract in a stock-trading app is precisely the participant the gambling framing worries about, and the industry’s response, self-surveillance regimes, contract-design standards, position limits, will be written under that spotlight. The cluster’s honest frame holds all three fronts at once: an exchange war for volume, a data war for institutional relevance, and a distribution war for the retail default, with the same November stress test scheduled for all three.
What to watch
The Schiff-Curtis whip count. The sports-ban bill is the one legislative vehicle that changes the revenue math rather than the compliance math. Its committee progress, and whether the exchange lobby’s fingerprints appear in the amendments, will reveal how far the political economy has already shifted.
Polymarket’s US re-entry mechanics. The QCEX license converting into live, fee-charging American operations is the event that would collapse the two rivals into one regulatory arena, and one comparable revenue table, for the first time. The valuation gap gets its verdict there.
The data products’ uptake. Signals feeds on terminals, Tradeweb integration metrics, the first hedge-fund letters citing prediction-market pricing as an input: the ICE thesis is falsifiable, and its evidence will be subscription revenue, not trading volume.
Midterm November. Record volumes are certain; the tests are operational integrity under peak load, insider-trading headlines under maximum scrutiny, and whether the state-law war produces an injunction that actually interrupts trading during the event. The category’s institutional era either survives its first American election as infrastructure, or it does not.
A final calibration for the cluster this piece opens: the honest bear case, stated without the valuations’ optimism. The category’s revenue concentrates in sports, the exact segment one live bipartisan bill would ban; its growth statistics are inflated by a World Cup and will meet a quieter 2027; its two leaders are burning venture capital on customer acquisition in a fee war that has already produced zero-fee promotional pairs; its insider-trading surface, markets on events their participants can influence or foreknow, is structurally worse than equities and has already generated congressional attention; and its institutional data thesis, however elegant, currently books a fraction of the revenue the trading does, meaning the valuations rest on the segment most exposed to the law and least proven under stress.
If the Schiff-Curtis framework passes, if a state supreme court cracks the preemption shield, or if November produces a manipulation scandal at scale, the $22 billion and $15 billion marks will be repriced as artifacts of the same exuberance that priced every previous financial-innovation wave at its legislative peak.
The bull case, argued above, is that the exchanges have already made the category too systemic to kill. The bear case is that Congress has killed systemic things before, and that betting on gambling’s legal normalization has broken sophisticated investors for a century. The cluster’s coverage will hold both, which is the only honest way to cover an industry whose defining trade is, fittingly, a binary contract on its own survival.
Frequently Asked Questions
What exactly did ICE invest in Polymarket?
A total commitment of $2 billion: $1 billion in October 2025, a further $600 million closing the agreement in March 2026, plus up to $40 million in secondary purchases from existing holders. The structure centers on global distribution rights to Polymarket’s event data alongside equity, and ICE launched institutional probability feeds, Polymarket Signals and Sentiment, in February, with a tokenization collaboration attached to the partnership.
How do Kalshi and Polymarket compare today?
Kalshi leads operationally: roughly $31.5 billion in June volume versus Polymarket’s $10.8 billion, fee revenue estimated between $850 million and $1.5 billion annualized, CFTC regulation, Robinhood distribution, and a Brazil expansion, at a $22 billion valuation after this spring’s billion-dollar raise. Polymarket holds the larger brand, crypto-native liquidity, the ICE partnership, and a CFTC-licensed US re-entry path via its QCEX acquisition, at a $9-10 billion valuation reaching toward $15 billion.
Why would the NYSE’s parent want a prediction market?
For the data, on the reading the deal structure supports. Prediction markets continuously price probabilities of discrete events, elections, rate decisions, geopolitical outcomes, a data product traditional exchanges have never generated, and ICE’s modern business is data distribution as much as trading. Chairman Jeffrey Sprecher framed the stake as a new layer of financial intelligence, and the February launch of institutional feeds executed exactly that thesis.
Are prediction markets actually profitable?
The leaders’ figures suggest strongly yes, with caveats. Kalshi’s revenue estimates range from $850 million in fees to $1.5 billion annualized against 994% growth into 2025; Polymarket only began charging taker fees this spring, with early annualized projections around $300 million; and Robinhood’s event-contract line reached $147 million in a quarter, exceeding its crypto revenue. Citizens Bank models the industry at a $3 billion run rate with a path toward $10 billion by 2030.
What is the legal threat, concretely?
Layered. At least seven 2026 bills target the category, led by the bipartisan Schiff-Curtis Prediction Markets Are Gambling Act, which would ban sports contracts on CFTC-regulated exchanges, the sector’s largest revenue source. State gaming regulators in a dozen-plus jurisdictions are litigating that event contracts are unlicensed wagers, with Nevada winning a temporary restraining order against Kalshi in March, while the CFTC backs federal preemption, the constitutional collision this publication’s earlier coverage maps.
Why does institutional money keep flowing despite the bills?
Because it is pricing a resolved endgame: preemption doctrine historically favors federal licenses, each acquisition shifts the political economy, banning bookies is easier than banning the NYSE’s data layer, and the probability-data thesis survives even adverse trading rulings, since data distributes globally regardless of where trading is licensed. The valuations treat the legal war as the moat’s construction cost, deterring entrants while incumbents entrench.
What is the joint VC fund the rivals backed?
A $35 million venture fund targeting the prediction-market sector, backed by the chief executives of both Kalshi and Polymarket, seeding infrastructure like analytics tooling and Bloomberg-style terminals for event markets. Rival operators jointly funding their category’s supply chain is the clearest signal that both view the sector’s growth, and its institutional legitimacy, as a shared asset that precedes their competition.
What should observers watch next?
Four things: the Schiff-Curtis bill’s committee progress, the one vehicle that changes revenue rather than compliance; Polymarket’s QCEX-licensed US operations going live, which puts both rivals in one arena; adoption evidence for the institutional data feeds, where the ICE thesis is falsifiable; and November’s midterms, the category’s biggest volume event colliding with its biggest legal fight, the first full-scale stress test of the institutional era. This is educational analysis, not investment advice.
Disclaimer: This article is for information and educational purposes only and does not constitute financial or investment advice. Valuations, volumes, and revenue figures are drawn from reporting and estimates that vary by source and change rapidly, and pending legislation and litigation described here are unresolved. Nothing here is a recommendation regarding any company, token, or contract. Always do your own research. Information is accurate as of July 24, 2026.
