The Border Inside the Contract

In June, Kalshi and Polymarket recorded more than $50 billion in wagers, including $5.69 billion on the FIFA World Cup final. Yet France blocked access to Polymarket over concerns about significant losses and manipulated trading. A prediction on the same event can be a regulated exchange product in the United States, prohibited gambling in France, or a loyalty-points game in Japan. The question does not change at the border. Everything that makes it tradable does.

Scale makes governance part of the product

An event contract looks simple. A venue defines a future outcome, participants trade claims tied to it, and the price condenses their research and beliefs into a probability. That price is useful only if regulators permit the contract and eligible traders can access it. The venue must also deter manipulation and resolve disputes credibly.

A small venue can treat weaknesses in those layers as product defects. At financial scale, they become public-market concerns. A disputed outcome can move substantial sums, while an insider can trade on information unavailable to others. A participant may even influence the event being priced. The venue is allocating gains and losses around elections, sports, corporate decisions, and other high-attention events.

World Cup trading pushed the two leading platforms past that threshold for the first time. Polymarket separately said annualized revenue exceeded $1 billion, while reported daily volume on its U.S. platform rose from about $50 million in mid-May to more than $200 million by June 20. As daily volume quadrupled in five weeks, Polymarket had to treat market integrity as an operating dependency rather than a policy appendix.

Neither venue’s volume proves that the market is manipulative, socially harmful, or poorly governed. It shows that failures can no longer be dismissed as the private consequences of a niche experiment. Under the CFTC’s proposed rules, the agency could bar contracts it finds contrary to the public interest or highly susceptible to manipulation. Exchanges must monitor trading, while regulators decide which contracts and users may reach the market.

Regulators did not suddenly discover prediction markets; billions of dollars changed the consequence of being wrong.

The borderless design depended on borders remaining irrelevant

Early prediction-market operators followed an economical model: list an event, attract opposing views, resolve the outcome, and let the price carry information across the internet. Distribution looked like a software problem, and settlement looked like a rules problem. Operators could treat the contract’s legal identity—derivative, gambling product, contest, or something else—as external until national authorities made that identity determine whether users could reach the interface.

Regulators changed access before they changed the underlying contract:

• 2024: French scrutiny led Polymarket to prevent French users from trading.
• November 2025: An amended CFTC designation created a route for Polymarket to operate through a fully regulated U.S. exchange structure.
• January 2026: Portugal ordered Polymarket to cease operations after more than €110 million in election-related volume.
• May 2026: Spain blocked Polymarket and Kalshi while investigating possible gambling-law violations.
• July 2026: France ordered access to Polymarket blocked over loss and manipulation concerns.

National regulators have created several legal markets sharing a brand and a basic contract form. A contract available through a regulated U.S. exchange can disappear behind a French block. For a distributor, the user’s location now narrows which venue it can offer before liquidity enters the decision.

Japanese startups Miraima and Poyp use loyalty points convertible to gift cards to navigate strict anti-gambling rules. The adaptation preserves the forecasting ritual while altering the money layer. Better distribution could not solve the legal problem; the startups needed a different product.

Polymarket appeared in 17 articles during the earlier 2024 period and 112 during the later 2026 period. Consumer framing fell from 41.2% to 17.9%, while enterprise framing rose from zero to 17%. Polymarket did not stop serving consumers, but reporters increasingly treated it as an exchange operator rather than only a consumer application.

A binary contract creates a nonbinary settlement problem

A binary interface hides layers of judgment. “Yes” and “No” work only after someone defines the event precisely, selects acceptable evidence, decides how to handle ambiguity, and makes the ruling credible to traders who may lose money because of it.

Polymarket uses Risk Labs’ Optimistic Oracle to decide roughly 200,000 difficult-call bets each month. A dispute over whether someone had uttered a single syllable showed how quickly semantic ambiguity becomes financial infrastructure. At this volume, Risk Labs and Polymarket must repeatedly decide how language, evidence, timing, and authority affect settlement.

Polymarket’s insider rules bar bets based on stolen confidential information, illegal tips, or outcomes a user can influence as an insider. Kalshi plans to require employment disclosures from users seeking to trade certain markets linked to material nonpublic information. Publishing a rule does not guarantee enforcement, but the rules require each venue to classify participants and conduct.

Sources also reported that the CFTC began an extensive investigation of Polymarket earlier in 2026. The agency did not confirm that account, so it remains a reported investigation rather than an established regulatory action.

Traders can use prices to estimate probabilities. Regulators still decide which contracts institutions may offer, while venue staff and oracle operators decide which evidence settles an ambiguous outcome and which conduct disqualifies a user.

Liquidity compounds only when distributors trust the venue

Event contracts are easy to describe and difficult to industrialize. A credible venue needs participants on both sides, dependable deposits and withdrawals, trusted resolution, integrity controls, and distribution partners willing to expose the product to their own customers. Distributors route activity only to venues they can defend. Their orders deepen liquidity, improve execution, and attract more traders.

When a broker compares two venues listing the same event, the question itself offers little basis for choosing. The broker instead asks whether each venue is authorized, liquid, able to settle disputes, monitoring manipulation, and permitted to serve its customers. Those checks drive prediction-market platformization: competitors can copy a question far more easily than they can win the broker’s order flow.

ADI Predictstreet launched with tiny trading volumes and user-reported withdrawal bugs. Its attempt to challenge Kalshi and Polymarket showed that listing contracts is only the visible fraction of venue construction. Traders experience software, confidence, and liquidity as one product.

Robinhood shifted away from Kalshi as it sought to diversify prediction-market providers and planned to route some World Cup bets to Rothera. A large distributor does more than deliver customers. It can choose which venue receives order flow, develops depth, and converts regulatory approval into a commercial advantage.

OpenAI, meanwhile, struck a deal to show Kalshi’s World Cup market data in ChatGPT search results. Prediction markets produce information as well as trades, so prices that travel through brokerage, social, or search interfaces gain reach that a standalone venue cannot reproduce.

That reach does not confer approval. Polymarket cannot offer margin trading in the United States without CFTC and NFA approval, while Robinhood can shift demand among providers it is willing to carry. A venue that loses eligibility can lose order flow without losing the underlying question.

The rivalry has moved from contracts to legitimacy

Kalshi and Polymarket still compete for traders, events, liquidity, and distribution. Each also wants regulators and distributors to accept its legal description of the category. The venue that wins that argument can enter a country, secure a broker, offer margin, or operate through a formally designated exchange.

Sources have described the companies as lobbying regulators, attempting to derail each other’s deals, and conducting influence campaigns. Those accounts remain allegations, not established facts. But when regulators authorize one venue and block another, the approved venue can gain users and distributors without changing the contract on the screen.

Both platforms have also partnered with social-media accounts purporting to be breaking-news reporters. The partnerships put market distribution inside an information environment capable of moving attention toward the events being traded. A feedback loop among headlines, attention, prices, and promotion can make the information channel part of the integrity problem.

The rivals now use tactics familiar from financial exchanges: listings, rules, trust, and access to participants. A competitor can reproduce the Yes-or-No wording, but it cannot inherit a rival’s designation, surveillance record, or broker relationships.

Regulation can enlarge the market it fragments

Polymarket’s amended CFTC designation gave it a route to reopen in the United States through a fully regulated exchange structure. A broker or counterparty evaluating the U.S. product can now point to that designation. Formal authorization does not guarantee distribution, but it gives partners a compliance case for carrying the venue.

Traders still pushed June wagers past $50 billion under those obligations. Their appetite is outrunning the institutional assumptions under which the category first grew.

Prediction markets can distribute risk, reward participants who investigate overlooked evidence, and compress dispersed knowledge into prices that others can use. None of that makes every contract socially valuable, every price reliable, or every restriction wise. A probability is an informational output, not a decision about what society should allow, prioritize, or do.

France blocked Polymarket over losses and manipulation concerns. The CFTC gave Polymarket a route through a designated U.S. exchange. Miraima and Poyp changed the money layer to comply with Japan’s anti-gambling rules.

A market can record $50 billion in a month and still stop at a national order, even though every contract ends in Yes or No. Around that binary outcome now stand a French blocking order, a CFTC designation, an employment disclosure, an oracle dispute, and a distributor’s routing table. The border has moved inside the contract.