Table of Contents
Kalshi is now valued at $22 billion, and Bernstein expects the prediction-market sector to reach $1 trillion in trading volume by 2030.
The major platforms are American and run at scale only in the US. What is not contained to the US is the regulatory response. Europe is already deciding how to treat a category its own operators have barely entered. Gibraltar has licensed Predictstreet as its first regulated operator and Malta is drafting a framework, while France, Germany, the Netherlands, and most of continental Europe have blocked the major US platforms or threatened fines. The UK treats them as a licensable category. The decisions facing European operators have to be made now.
This guide is for operators trying to read that map: what prediction markets are, why the demand is real, what the regulation looks like, and what the decision looks like when the infrastructure is American, but the players are everywhere.
How Prediction Markets Work
A prediction market is a place to “bet” on real-world events using binary-outcome contracts. Users buy contracts tied to a specific future event. Each contract pays $1 if the outcome is yes and $0 if it is no. Because the maximum payout is fixed, the price of a contract lies between 0 and 1 and can be interpreted directly as a probability.
Take the 2024 US presidential election. Polymarket’s contract on a Trump victory traded above $0.60, which meant the market priced his chances at around 60%. That one market handled $3.6 billion in volume. When the election resolved, the winning bets were paid from the pool of losing ones.
Two models share the category today. One is dominant in the US. The other is emerging in Europe.
The trading model
Polymarket and Kalshi run on the trading model. Prices move continuously, so users can buy and sell contracts with each other before the event resolves, just as they could close a stock position before a company reports earnings. The platform takes a transaction fee on each trade rather than a bookmaker’s margin. The operator does not take the other side of any bet.
That structure is why the model is legally classified as a financial product rather than gambling. In the US, prediction markets are regulated by the Commodity Futures Trading Commission as derivative event contracts. The regulatory route exists, has been tested in court, and supports markets at scale. No European regulator treats this category as a financial product the way the CFTC does.
The betting model
A second model is now emerging from B2B vendors targeting iGaming operators. SOFTSWISS launched its Prediction Markets product in April 2026, adapting the format to a sportsbook-style mechanic.
Users buy contracts but cannot sell them. The operator sets the prices, bears the risk, and earns a margin built into the odds, exactly like a sportsbook. Because contracts cannot be exited before resolution, they are wagers on future events. That makes them gambling in most jurisdictions’ legal definitions.
The practical consequence is that the betting model can often run on an operator’s existing sportsbook licence in jurisdictions where event betting beyond sports is permitted, like the UK. In more restrictive markets like Germany and France, sportsbook licences cover only sports events, so a separate licence is needed. The trading model is harder. It requires either a US-style financial services framework or a UK-style betting intermediary licence, and neither route is available in most of continental Europe.
How the two models compare
| Trading model | Betting model | |
|---|---|---|
| Used by | Polymarket, Kalshi | B2B vendors |
| User action | Buy and sell contracts with each other | Place a bet, held to settlement |
| Pricing | Set by the market | Set by the operator |
| Operator role | Match trades, take a fee | Set prices, carry the risk |
| Revenue | Transaction fees | Margin in the odds |
| Legal category | Financial product | Gambling |
| Licence needed | Financial-services (CFTC in US) | Existing sportsbook licence |
Trading models give you a price that reflects what thousands of users actually believe, which is what made Polymarket’s Trump number interesting. But they need sufficient liquidity to function, and a regulator that recognises the financial product classification.
Betting models are simpler to run, integrate naturally into existing iGaming stacks, and may fit within licenses operators already hold. They lose the wisdom-of-crowds signal that makes prediction markets different from regular betting in the first place.
Why Prediction Markets Are Growing in 2025 and 2026
Two forces are pulling demand at once: retail volume that looks structurally different from sports betting, and serious institutional money now backing the major platforms in the US.
Retail volume is the most visible. Polymarket’s monthly notional volume reached $10.3 billion in April 2026, nearly triple the $3.8 billion it ran in April 2025. Engagement on these platforms looks structurally different from traditional betting. Yield Sec’s monitoring, cited in iGaming Business’s Europe feature, found users returning to check their prediction-market positions roughly 15 times per hour on the first day after placing a trade. That is engagement closer to retail trading apps than to sportsbooks, and it is what makes the category sticky.
The institutional side is newer. Intercontinental Exchange, parent of the New York Stock Exchange, committed $1.6 billion to Polymarket across two tranches by April 2026, with Polymarket in talks to raise a further $400 million at a $15 billion valuation. Charles Schwab and Citadel Securities have publicly said they are evaluating moves into the sector. CME Group already lists event contracts, while Nasdaq and Cboe have filed to launch their own.
Bernstein expects the sector to reach $1 trillion in trading volume by 2030, against $51 billion in 2025 and roughly $240 billion projected for 2026, an implied compound growth rate of around 80% per year.
Prediction Markets Regulation in the US
In the US, prediction markets are regulated as financial products rather than gambling. Kalshi and Polymarket are licensed by the Commodity Futures Trading Commission under the Commodity Exchange Act, the same federal statute that governs derivatives and futures trading. Their contracts are classified as event-based derivatives, which, in principle, place them outside state gambling law.
In practice, the split between federal and state authority is being argued out in court. The most important ruling so far came in April 2026, when the Third Circuit decided in KalshiEx v. Flaherty that federal commodities law preempts state gambling regulation of CFTC-licensed prediction markets. It was the first federal appellate ruling on the question and is widely expected to reach the Supreme Court before the end of the year.
Underneath the legal framework, the CFTC is also working on integrity. Its March 2026 advisory and accompanying Advance Notice of Proposed Rulemaking opened a formal consultation on event-contract regulation, including how to handle insider trading and market manipulation. The concerns are not abstract. In April 2026, a US Army Special Forces soldier was indicted for making over $400,000 on Polymarket using classified information about a planned military operation, and Kalshi has flagged more than 400 suspicious trades in 2026, more than double its total for all of 2025. Enforcement is now active, even if the framework around it is still being written.
For operators looking at any other market, this is the reference case. It is the only jurisdiction where prediction markets operate at scale, and even there, the rules are still being settled.
Prediction Markets in Europe: A Fragmented Regulatory Picture
Europe is moving in three directions at once. The largest cluster of European regulators has blocked the major US platforms. The UK is treating them as a category that already fits its existing licence regime. Gibraltar and Malta are opening doors. No European country has yet matched the US approach of regulating prediction markets as financial products.
The blocking position is the loudest. France’s Autorité Nationale des Jeux ordered Polymarket to geoblock French users in November 2024 and followed up in February 2026 with a broader warning that prediction markets carry amplified addictive characteristics relative to licensed gambling. The Netherlands’ Kansspelautoriteit went further in February 2026, ordering Polymarket out of the country under threat of an €840,000 fine. Germany, Belgium, Portugal, Switzerland, Romania, Greece, Poland, and Cyprus have all blocked access to Polymarket on the same grounds. The shared position across these markets is that prediction markets meet the legal definition of gambling and require a licence that the platforms do not hold.
The UK is the closest functional analogue to a regulated prediction market in Europe. In a February 2026 blog post, the UK Gambling Commission set out that prediction markets fall within the definition of a Betting Intermediary under British law, the same category that covers Betfair and Smarkets. Polymarket and Kalshi both geoblock UK users, but the licensed exchange model that runs the same matching mechanism has existed in Britain since 2000. The infrastructure already exists. Only the US-named platforms are excluded.
The opening doors are Gibraltar and Malta. Gibraltar licensed ADI Predictstreet in April 2026 as Europe’s first regulated prediction-market operator, and FIFA made it the Official Prediction Market Partner of the 2026 World Cup. Malta is exploring its own framework. Both jurisdictions are running the same playbook they used for iGaming and crypto: become the regulated European home for a new category before the larger markets settle their own rules.
One structural point sits underneath all of this. There is no European CFTC. Financial regulators handle derivatives, gambling regulators handle betting, and prediction markets fall in the seam. No supranational authority can resolve the question the way the US is trying to, which is why the European picture will stay fragmented for the medium term.
Will Prediction Markets Cannibalise Sports Betting?
The cannibalisation question has a clear answer in the only market where it has been tested at scale. Yes, prediction markets are taking sports betting volume. The question is how much and from whom.
Kalshi’s 2025 financials make the case starkly. Of Kalshi’s $263.5 million in fee revenue in 2025, 89% came from sports, with the sports share exceeding 90% in the final four months of the year. During the September-November 2025 NFL season, Kalshi held roughly 3% of the US sports betting market, ranking it approximately seventh among US sports betting operators by revenue, ahead of Rush Street Interactive and theScore Bet. On Super Bowl Sunday 2026 alone, Kalshi processed $871 million in trading volume, most of it tied to the game itself.
The major US sportsbooks did not sit and watch. DraftKings and FanDuel both launched their own prediction-market products in late 2025 and early 2026, including in states where they hold no sports betting licence. The competitive logic was straightforward: a prediction market with sports contracts is a way to reach the same demand under a federal regulatory framework that does not require a state-by-state sports betting licence.
For European operators, the lesson is not whether cannibalisation happens. It does. The lesson is that ignoring it for long is not an option taken by serious operators in markets where prediction markets are available. The largest US sportsbooks treated the category as a competitive threat that required matching, not avoiding.
Models Available in Europe Today
For most European operators, there are really two practical options today. Wait, or partner with a B2B vendor.
For most mid-sized operators, the B2B vendor route is currently the most practical path. Early B2B products are already appearing, with fixed-odds prediction markets as the clearest model. In this setup, the operator prices the market, carries the risk, and earns from the margin built into the odds. In markets with broader betting or exchange frameworks, such as the UK and potentially Malta, the route may be easier to assess. In stricter markets such as Germany, where sports betting permissions are limited to sports events, non-sport prediction markets face a much harder licensing path.
A peer-to-peer exchange model is possible in principle in the UK under the Betting Intermediary framework, where the operator brings betting parties together without carrying liability for their bets. Gibraltar also appears to support this route, having approved Predictstreet as a regulated prediction-market operator. Both routes are more complex than fixed-odds betting and require direct regulator or licence-specific review.
Two other routes exist, but they are unlikely to be realistic first steps for most mid-sized operators. Building proprietary infrastructure through a Gibraltar route, or a future Malta route if one becomes available, is viable for specialist operators but likely requires more capital, legal work, and regulatory engagement than a vendor-led product. For UK-facing operators, the long-established betting exchange model is the closest reference point, though replicating that model elsewhere requires a local legal basis for exchange-style or intermediary betting.
For most operators in 2026, the B2B vendor route is the most realistic way to test prediction markets without building exchange infrastructure or leading a major licensing project from scratch.
What to Watch Over the Next 12-24 Months
Three things will shape the European operator decision more than anything else.
Predictstreet at the World Cup. Predictstreet is the only operating regulated prediction market in Europe. The 2026 FIFA World Cup is its first real test. If it runs cleanly and the numbers hold up, the Gibraltar route gets validated and Malta will move faster on its own framework. If it stumbles, the European model loses momentum for a year or more.
The Tipico aftermath. On 16 April 2026, the European Court of Justice ruled in Case C-440/23 that EU Member States can prohibit online gambling services from operators licensed elsewhere in the EU, and can impose civil consequences including contract nullity and restitution of player losses. This applies directly to prediction markets. National regulators now have stronger legal grounds to act against offshore prediction-market operators serving their residents. Watch which Member States actually use these powers, and whether restitution claims start landing against the major offshore platforms.
The first major European operator move. Flutter has launched FanDuel Predicts in the US and already operates Betfair Exchange in Europe, but no European-licensed operator has launched a product specifically positioned as a prediction market for European markets. The first one that does will set the template, the price point, and the competitive pressure for everyone else. Watch for acquisitions, B2B partnerships, and European-market product launches from the larger UK, Maltese, and pan-European operators.
Should You Do Prediction Markets?
Five questions decide it.
Where are we launching? Europe is not one market. The UK, Gibraltar, and Malta have licence routes. France, Germany, the Netherlands, and most of continental Europe have blocks. These jurisdictions cannot share a strategy.
What can we legally offer? Sports usually fits existing betting permissions. Politics, elections, finance, and entertainment usually do not. In most markets, expanding event scope means a new licence or no licence at all.
Trading or betting? Trading needs liquidity, a different licence, and customers who behave like stock traders. Betting fits an existing sportsbook stack but puts the pricing and risk on you.
New revenue or cannibalisation? Some of the new audience is genuinely new. Some is the existing sportsbook customer paying you in a different way. Know which before you launch.
Can we run it cleanly? Event settlement, responsible gambling, advertising limits, suspicious trading, insider-information risk, and player complaints all need owners before launch. The integrity issues the CFTC is dealing with right now apply in every other market too.
Operators with clear answers to all five are ready to move.
Reading the Map
Prediction markets are already operating. The US infrastructure runs at scale, and European regulators are choosing positions through 2026 and 2027.
For most operators in Europe, the practical decision in 2026 is between waiting and partnering with a B2B vendor. The five questions in the previous section are the way to make either choice clearly.
Operators who decide now set the terms for everyone else.
FAQ
What is the difference between a prediction market and a betting exchange?
A prediction market lets users buy and sell binary outcome contracts on real-world events. A betting exchange like Betfair lets users match bets against each other on sports outcomes. The mechanism is essentially the same. The difference is event scope: betting exchanges cover sports, prediction markets cover anything verifiable.
Are prediction markets legal in the UK?
Yes, with the right licence. The UK Gambling Commission has confirmed that prediction markets fall within the definition of a Betting Intermediary, the same category that covers Betfair and Smarkets. Polymarket and Kalshi do not hold UK licences and geoblock UK users.
Are prediction markets legal in the European Union?
Mostly no. France, Germany, the Netherlands, Belgium, Portugal, Romania, Greece, Poland, and Cyprus have all blocked the major US platforms. No EU member state has a dedicated licence framework yet, though Malta is working on one. There is no EU-wide regulator for the category.
How do prediction markets make money?
Trading-model platforms like Polymarket and Kalshi charge transaction fees on each trade between users. Betting-model platforms set the odds and earn from the margin built into the prices, like a sportsbook. Neither model relies on users losing for the platform to earn.
