A prediction market is an exchange where people trade contracts on whether a future event will happen: a typical contract pays $1 if it does and $0 if it doesn't, so its price, somewhere between $0 and $1, is read as the market's estimate of the probability. The exchange matches traders who disagree and usually charges fees; on exchanges regulated in the US, it doesn't take a side itself.
The US regulator for these markets, the Commodity Futures Trading Commission (CFTC), says prediction markets offer products "designed to help the public forecast, plan for, hedge, and even harness perceptions of future events." Each product is an event contract: a derivative whose payoff depends on whether a specified event happens. Most are yes-or-no questions with a fixed payout, usually $1, and an expiry date.
A worked example. Suppose a contract asks whether an event will happen by a set date, and Yes trades at 70¢. The CFTC uses the same number in its own explainer: a 70¢ price means the market puts the chance at about 70%. No then costs about 30¢, because one Yes and one No together always pay exactly $1.
If you buy 10 Yes contracts at 70¢, you pay 10 × $0.70 = $7.00. Your counterparty is another trader, for example someone who buys the 10 matching No contracts at 30¢ for 10 × $0.30 = $3.00. Together that is $10.00, exactly what the 10 contracts will pay out. If the event happens, you receive 10 × $1 = $10.00, a $3.00 gain, and the No buyer's $3.00 is gone; if it doesn't, the No buyer receives the $10.00 and your $7.00 is gone. Before fees, one side's gain is the other side's loss.
What the exchange does. The platform lists the questions, keeps an order book of bids and offers, matches buyers with sellers and settles the contracts when the result is known. It doesn't set the price: prices move as traders buy and sell. The CFTC says the exchanges it regulates "do not take a side of the trade" and "are not competing against you"; Kalshi, for example, says it earns transaction fees and takes no position.
Other designs. Some markets pay in proportion to a number, such as a candidate's share of the vote, instead of $1 or $0, and a question with several possible answers is usually split into one Yes/No contract per answer. This page focuses on the common $1/$0 contract.
Platforms differ in the details, but every trade goes through the same four stages. The numbers below use Kalshi's published taker fee.
For the screens, order types and fees on Polymarket, see how Polymarket works, how to buy on Polymarket and how to sell on Polymarket; for Kalshi, see how Kalshi works. The fee calculator works out fees at any price.
Because a contract pays $1 or $0, a trader who thinks the chance is 45% values a Yes contract at about 45¢. Buyers step in while the price is below what they believe and sellers while it is above, so the price settles near the traders' average estimate. Polymarket's docs say the price "directly represents the market's belief in the probability" of the outcome, and Kalshi's help center reads it the same way: Yes at 70¢ and No at 30¢ mean a 70% consensus.
That reading is approximate. Economists Wolfers and Zitzewitz found that prices are usually close to traders' average belief, typically within a percentage point, with the largest gaps near $0 and $1. The number shown on a market page may also not be a price you can trade at: on Polymarket a 34¢ bid and a 40¢ ask show as 37%, while buying costs at least 40¢ plus any fee. In the Kalshi example above, the trade only has a positive expected value if the true chance is above $41.68 ÷ $100.00 = 41.68%. See prices are probabilities: how to read prediction market odds for conversions to betting odds and more worked examples.
Versus polls. A poll asks a sample of people what they intend or prefer; a prediction market asks traders to put money on what they expect to happen, and the price pools what they know. Wolfers and Zitzewitz describe markets as able to "aggregate disperse information into efficient forecasts". A poll measures opinion at one moment; a market forecasts an outcome, and it can be wrong too.
Versus stocks. A share of stock is an ownership stake in a company with no expiry date. An event contract is a derivative that expires on a set date and pays a fixed amount, $1 or $0, depending only on whether the event happened. In the US, event contracts are overseen by the CFTC under commodities and derivatives law, not by the securities regulator.
Versus sportsbooks. A bookmaker sets the odds itself, takes the other side of your bet and builds its margin into the odds. An exchange instead matches customers against each other and charges a fee. A Harvard Business School study of the UK market found that Betfair's exchange margin (the over-round) was typically under 3%, against about 12% at bookmakers. Prediction-market exchanges follow the exchange model, which is why the UK regulator compares them to betting exchanges (see below). Critics dispute Kalshi's framing that it isn't acting as the house, pointing to affiliated market makers and fees, and argue the product works like sports betting.
How they compare:
| Type | Your counterparty | Who sets the price | How you win or lose | Who regulates it (examples) |
|---|---|---|---|---|
| Prediction market (US-regulated exchange) | Another trader on the opposite side; the exchange only matches you | Traders, through bids and offers on the order book | Each contract pays $1 if your side is right and $0 if not; fees on top | The CFTC at federal level; many states dispute sports contracts |
| Opinion poll | None: nothing is bought or sold | No price: the pollster reports the share giving each answer | Nothing is staked | Not a traded financial product |
| Stock | Another investor, matched by the exchange | Buyers and sellers | Value rises or falls with the share price; no fixed payout or end date | Securities regulators (the SEC in the US), not the CFTC |
| Sportsbook (bookmaker) | The bookmaker itself | The bookmaker, with its margin built into the odds | Win the payout at the odds you took, or lose your stake | US: state gaming regulators, licensed state by state |
| Betting exchange (e.g. Betfair) | Another bettor: one backs, the other lays | Bettors' offers | Win at the matched odds or lose your stake; commission on winnings | UK: the Gambling Commission |
There is no single answer: it depends on the regulator and the country. PolyKal reports how each one classifies them and doesn't take a side.
US federal level. The CFTC treats event contracts as derivatives, "typically structured as swaps", traded on exchanges it designates (designated contract markets). Exchanges list new contracts by certifying them themselves, and the CFTC can bar contracts involving terrorism, assassination, war, gaming or unlawful activity if it finds them contrary to the public interest.
The 2026 CFTC proposal. A 2024 CFTC proposal to ban political and some sports contracts as "gaming" was withdrawn under Chairman Michael Selig, announced in late January and formally withdrawn in February 2026. On June 10, 2026 the CFTC proposed a new framework: it defines "gaming", treats elections and awards as contests rather than games, presumes that contracts on scores, spreads and statistics in professional and college sports are allowed, and presumes that contracts on injuries, officiating, "first action" events and pre-college sports are not. The rule is a proposal; public comments closed on July 27, 2026. The law firm DLA Piper reports that 44 states jointly disputed the CFTC's authority in their comments.
US states. Many states say sports event contracts are sports betting that needs a state license, and have sent cease-and-desist letters or gone to court. The courts disagree. In April 2026 the Third Circuit sided with Kalshi in a New Jersey case, finding state gaming law likely preempted; on August 28, 2026 the Ninth Circuit ruled 3–0 for Nevada that sports event contracts are not "swaps" and that Nevada law is likely not preempted. DLA Piper counts 19 states in proceedings, and the CFTC itself has sued several states over the question. In early September 2026 New Jersey asked the Supreme Court to decide, as did Crypto.com and Robinhood after losses in the Ninth Circuit; we found no decision to hear the case as of September 24, 2026.
Outside the US. Regulators elsewhere generally treat prediction markets as gambling. The UK Gambling Commission wrote in February 2026 that their "core aspects are akin to what in the UK would be described as a 'Betting Exchange'", so operators need a gambling license to serve Great Britain, while financial products fall to the Financial Conduct Authority. Nine European gambling regulators issued a joint statement in June 2026, and authorities in countries including France, Italy, Spain, Brazil and India have ordered blocks of Polymarket, Kalshi or both.
As of September 2026 the main platforms fall into a few groups: exchanges regulated by the CFTC in the US, brokers and apps that pass customers' orders to those exchanges, an international crypto-based platform, and play-money or academic markets. The table lists only platforms our research covers.
| Platform | Type | Who can use it | Money |
|---|---|---|---|
| Kalshi | US exchange regulated by the CFTC since 2020 | US users (several states dispute its sports contracts) and, since October 2025, users in many other countries; its member agreement lists 55 restricted countries | US dollars; international members can also deposit by debit card, wire or crypto |
| Polymarket (polymarket.com) | International crypto-based exchange on the Polygon blockchain; users hold their own wallets | Adults (18+) outside the US; its help center lists dozens of blocked countries, including the US, the UK and France | Crypto: pUSD, a token backed 1:1 by USDC |
| Polymarket US | Separate US exchange regulated by the CFTC; accounts aren't connected to polymarket.com | US residents after an identity check, in the iPhone app (an Android version is listed on Google Play, though Polymarket's help center still says it is coming soon); its App Store listing says trading is unavailable in Nevada | US dollars (debit card, bank transfer or wire) |
| Robinhood | Broker (futures commission merchant) that routes orders to exchanges, including Kalshi and ForecastEx | Robinhood's US customers | US dollars |
| ForecastEx (Interactive Brokers) | CFTC-registered exchange and clearinghouse owned by Interactive Brokers, launched in August 2024 | Interactive Brokers clients through its ForecastTrader platform; also reachable through Robinhood | US dollars |
| FanDuel Predicts (CME Group) | App offering contracts listed by CME Group, launched December 22, 2025 | Launched in five states (AL, AK, SC, ND, SD) with a wider rollout planned; sports contracts only where online sports betting isn't legal | US dollars |
| Manifold | Play-money market | New users get Ṁ1,000 of play money; anyone can create a market | Play money ("Mana") that can't be converted to cash |
Others include Crypto.com's CFTC-registered exchange (sports contracts since December 2024), Rothera, a Robinhood–Susquehanna venture that listed its first contracts in May 2026, PredictIt and the Iowa Electronic Markets. In the UK, Betfair runs a betting exchange under a Gambling Commission license. For the two biggest side by side, see Polymarket vs Kalshi compared, our Polymarket review and our Kalshi review.
Trading has grown fast, mostly in sports. Monthly volume on Kalshi and Polymarket combined, as counted by The Block and reported by the Pew Research Center, went from under $5 billion in September 2025 to about $24 billion in April 2026 and a peak of $53.0 billion in July 2026, during the World Cup: $40.1 billion on Kalshi and $12.89 billion on Polymarket and Polymarket US. In August 2026 it fell 14.5% to $45.33 billion, the first monthly drop in a year (The Block, September 2, 2026).
Counting methods differ. The Block and Pew count notional volume on the taker side only, valuing each contract at $1. TRM Labs counts both sides of each trade and reported more than $20 billion a month in January 2026, against about $1.2 billion a month in early 2025. Kalshi's own "annualized" figures (a rise from $52 billion to $178 billion in six months, self-published in May 2026) are extrapolations, and Robinhood's self-published figures count contracts (13.6 billion in Q2 2026), not dollars. A Columbia University study estimated that about 25% of Polymarket's historical volume was likely wash trading, meaning trades with no real change of position. These figures can't be compared directly.
Where the money goes. From July 2024 to April 2026, 80% of Kalshi's volume was sports and 4% politics; on Polymarket's international site it was 39% sports and 32% politics (Pew). Pew found that prediction-market sports volume in the first quarter of 2026 was about on par with the amount wagered at US legal sportsbooks. Investors valued Kalshi at $22 billion in May 2026 and Polymarket at about $21 billion in a round reported at the end of August 2026.
Across many markets, often quite accurate, but not always. Compared with 964 polls in five US presidential elections from 1988 to 2004, the Iowa Electronic Markets were closer to the final result 74% of the time (Berg, Nelson and Rietz, 2008), and a 2026 study of Kalshi found prices broadly match how often events happen, getting more accurate as expiry nears. But a working paper on more than 2,500 markets in the 2024 US election found identical contracts priced differently across exchanges, and the platforms' own accuracy pages don't disclose their full method. See how to read prediction market odds for the evidence in detail.
These problems are documented; the sources are listed at the end of the page.
It depends on your country and, in the US, your state, and the rules change often. Kalshi's member agreement lists 55 restricted countries, polymarket.com blocks the US and dozens of other countries, and governments have ordered blocks that go beyond both platforms' own lists. See where Polymarket and Kalshi are available for the current, dated list.
Polymarket's Terms of Use prohibit using a VPN or similar tool to get around these restrictions, and doing so can also break local law. PolyKal doesn't help anyone get around them.
A prediction market is an exchange where people buy and sell contracts on future events, such as elections, economic data or sports results. A typical contract pays $1 if the event happens and $0 if it doesn't, so its price between $0 and $1 is read as the market's estimate of the probability: 70¢ means about 70%.
You buy Yes or No contracts from other traders through the exchange's order book, at prices set by supply and demand. Before the end you can usually sell at the current bid; at the end, the market resolves under its written rules, and each winning contract pays $1 while each losing one pays $0. The exchange usually charges fees and, on US-regulated venues, doesn't take a side.
It depends on where you are, and this isn't legal advice. In the US, Kalshi and Polymarket US are exchanges regulated by the CFTC, but many states argue their sports contracts are unlicensed betting, federal appeals courts have split, and Supreme Court petitions were filed in September 2026. Many other countries treat prediction markets as gambling and have blocked some sites. See where Polymarket and Kalshi are available.
Regulators disagree. The US CFTC treats event contracts as derivatives and in June 2026 proposed rules that treat elections as contests rather than games; the Ninth Circuit found in August 2026 that sports event contracts are not "swaps" and that Nevada's gaming law is likely not preempted; and the UK Gambling Commission says prediction markets are akin to betting exchanges. Whatever the label, you can lose the full amount you put in.
Often fairly accurate across many markets, especially close to the end: the Iowa Electronic Markets were closer to the result than polls 74% of the time across five presidential elections, and prices on Kalshi broadly match outcomes. But cheap longshots win less often than their price suggests, the same event can trade at different prices on different sites, and a single result can't prove a market right or wrong. See how to read prediction market odds.
Kalshi is a US exchange regulated by the CFTC since 2020; it uses US dollars and also accepts users in many other countries. Polymarket's main site, polymarket.com, is a crypto-based platform using pUSD, a USDC-backed token, and doesn't accept US residents, who use the separate, CFTC-regulated Polymarket US app in dollars. Fees differ too: Kalshi's taker fee is 0.07 × contracts × price × (1 − price), and polymarket.com's rate runs from 0 to 0.07 depending on the category. In August 2026 Kalshi traded $37.17 billion, against $8.16 billion for Polymarket and Polymarket US (The Block). See Polymarket vs Kalshi compared.
Some traders do, but on average traders lose. Prices already reflect what the crowd knows, so you only gain if you judge the odds better than everyone else; fees and bid-ask spreads cost money on every trade; and many cheap, small-probability contracts lose. A study of Kalshi found an average return of about −20% on contracts and about −31% for takers, and a working paper on Polymarket found that fewer than one-third of traders make a profit. PolyKal doesn't give trading advice.
Traders do. The price is where buyers and sellers agree to trade on the order book, so it moves with every trade; the exchange lists the market, matches orders and settles the result. That is the main difference from a sportsbook, where the bookmaker sets the odds and takes the other side.
Not on a standard $1/$0 contract that you buy: the most you can lose is what you paid plus fees, because the contract can't be worth less than $0. Polymarket and Crypto.com, for example, describe their contracts as fully backed by collateral. You can, however, lose all of it.
By trading volume, Kalshi was well ahead of Polymarket in July and August 2026: in August The Block counted $37.17 billion on Kalshi and $8.16 billion on Polymarket and Polymarket US. In March 2026 Pew found the two roughly level, with about half of total volume each. Rankings depend on how volume is counted.
Every fact on this page comes from the sources below, checked on September 24, 2026. Rules, court cases, fees and volumes change often, so figures are as of September 2026, and we re-verify this page quarterly. Statistics that platforms publish about themselves are labeled as self-published, and where sources disagree we say so or leave the point out. See our methodology.
Regulators and courts: CFTC: Understanding Prediction Markets and Event Contracts, CFTC: Contracts & Products, CFTC: Kalshi designated as a contract market (2020), CFTC: order against Polymarket (2022), CFTC: complaint against Intrade (2012), CFTC Letter 25-20 (PredictIt, 2025), Federal Register: Prediction Markets, advance notice (March 2026), CFTC: proposed rule (June 2026), Federal Register: proposed rule (June 2026), D.D.C. decision in KalshiEX v. CFTC (2024), D.C. Circuit order in KalshiEX v. CFTC (2024), Third Circuit opinion in KalshiEX v. Flaherty (2026), Ninth Circuit opinion in KalshiEX v. Assad (2026), US Justice Department charge (April 2026), UK Gambling Commission: prediction markets (February 2026), German gambling regulator (GGL): joint statement of European regulators (June 2026), ANJ (France): Polymarket blocking order (July 2026), DGOJ (Spain): proceedings against Polymarket and Kalshi (May 2026), Iowa Electronic Markets FAQ.
Platforms: Kalshi: What Is Kalshi?, Kalshi help: How are prices determined?, Kalshi fee schedule, Kalshi member agreement, Kalshi help: trading from outside the US, Kalshi: $1 billion raise (self-published), Polymarket help: What is Polymarket?, Polymarket help: How Are Prices Calculated?, Polymarket help: How Are Markets Clarified?, Polymarket help: Can I Sell Early?, Polymarket help: Does Polymarket Have Trading Limits?, Polymarket help: Geographic Restrictions, Polymarket help: Website vs App, Polymarket help: Is My Money Safe?, Polymarket docs: prices and order book, Polymarket docs: fees, Polymarket docs: resolution, Polymarket docs: how positions work, Polymarket docs: pUSD, Polymarket Terms of Use, Polymarket US docs: what is Polymarket US, Polymarket US docs: deposit methods, App Store: Polymarket, Polymarket: market-integrity rules (March 2026), Robinhood newsroom (September 2026), Robinhood: prediction markets hub (March 2025), ForecastEx launch release, CME Group: FanDuel Predicts, Crypto.com help: sports event trading, Crypto.com: sports launch (December 2024), Manifold: About, Betfair: What is a betting exchange?.
Research: Wolfers & Zitzewitz, "Prediction Markets" (2004), Wolfers & Zitzewitz, "Interpreting Prediction Market Prices as Probabilities" (2006), Arrow et al., "The Promise of Prediction Markets" (2008), Berg, Nelson & Rietz, "Prediction market accuracy in the long run" (2008), Bürgi, Deng & Whelan on Kalshi (VoxEU, 2026), Clinton & Huang on the 2024 election markets (working paper), Reichenbach & Walther on Polymarket (working paper), Rothschild & Sethi on Intrade, Tsang & Yang on Polymarket's 2024 market (preprint), Casadesus-Masanell & Campbell on Betfair (Harvard Business School), Rose-Berman, "Kalshi's Favorite Lie" (opinion).
Data and reporting: Pew Research Center (May 2026), Pew Research Center (September 2026), The Block: August 2026 volume, TRM Labs on monthly volume, DLA Piper tracker (September 2026), Ropes & Gray on the CFTC proposal, Corporate Compliance Insights on the withdrawn 2024 proposal, Insider Sport on the Supreme Court petitions, Decrypt on the 2024 election market, Decrypt on the Columbia wash-trading study, NBC News on the French trader, NPR on Kalshi's insider cases, CoinDesk on the Ukraine minerals market, Decrypt on the Zelenskyy suit market, CasinoBeats on the Khamenei market, CoinDesk on the June 2026 security incident, TechCrunch on Kalshi's valuation, Bloomberg on Polymarket's valuation, PBS NewsHour on the CFTC suing states (April 2026), Lefosse on Brazil's rules (April 2026), ThePrint on India's blocking orders (May 2026), Punto Informatico on Italy's blacklist (July 2026).
History: Quartz on Intrade's shutdown (2013), CCH summary of the PredictIt ruling (2025), Wikipedia: PredictIt (secondary), Yahoo Finance on the CFTC dropping its Kalshi appeal, PR Newswire: Polymarket acquires QCEX, CoinDesk on Kalshi's expansion to 140 countries, CoinDesk on the Polymarket US app launch, Yahoo Finance on Polymarket US opening to all iPhone users, DeFi Rate on Rothera (secondary), ICE announcement naming Polymarket's founding.
PolyKal is an independent research tool, not affiliated with Polymarket, Kalshi or any platform named here. This page explains what prediction markets are; it is not financial, legal or tax advice, and not an invitation to open an account or trade.
A prediction market is an exchange where people trade contracts on whether a future event will happen: a typical contract pays $1 if it does and $0 if it doesn't, so its price, somewhere between $0 and $1, is read as the market's estimate of the probability. The exchange matches traders who disagree and usually charges fees; on exchanges regulated in the US, it doesn't take a side itself.
The US regulator for these markets, the Commodity Futures Trading Commission (CFTC), says prediction markets offer products "designed to help the public forecast, plan for, hedge, and even harness perceptions of future events." Each product is an event contract: a derivative whose payoff depends on whether a specified event happens. Most are yes-or-no questions with a fixed payout, usually $1, and an expiry date.
A worked example. Suppose a contract asks whether an event will happen by a set date, and Yes trades at 70¢. The CFTC uses the same number in its own explainer: a 70¢ price means the market puts the chance at about 70%. No then costs about 30¢, because one Yes and one No together always pay exactly $1.
If you buy 10 Yes contracts at 70¢, you pay 10 × $0.70 = $7.00. Your counterparty is another trader, for example someone who buys the 10 matching No contracts at 30¢ for 10 × $0.30 = $3.00. Together that is $10.00, exactly what the 10 contracts will pay out. If the event happens, you receive 10 × $1 = $10.00, a $3.00 gain, and the No buyer's $3.00 is gone; if it doesn't, the No buyer receives the $10.00 and your $7.00 is gone. Before fees, one side's gain is the other side's loss.
What the exchange does. The platform lists the questions, keeps an order book of bids and offers, matches buyers with sellers and settles the contracts when the result is known. It doesn't set the price: prices move as traders buy and sell. The CFTC says the exchanges it regulates "do not take a side of the trade" and "are not competing against you"; Kalshi, for example, says it earns transaction fees and takes no position.
Other designs. Some markets pay in proportion to a number, such as a candidate's share of the vote, instead of $1 or $0, and a question with several possible answers is usually split into one Yes/No contract per answer. This page focuses on the common $1/$0 contract.
Platforms differ in the details, but every trade goes through the same four stages. The numbers below use Kalshi's published taker fee.
For the screens, order types and fees on Polymarket, see how Polymarket works, how to buy on Polymarket and how to sell on Polymarket; for Kalshi, see how Kalshi works. The fee calculator works out fees at any price.
Because a contract pays $1 or $0, a trader who thinks the chance is 45% values a Yes contract at about 45¢. Buyers step in while the price is below what they believe and sellers while it is above, so the price settles near the traders' average estimate. Polymarket's docs say the price "directly represents the market's belief in the probability" of the outcome, and Kalshi's help center reads it the same way: Yes at 70¢ and No at 30¢ mean a 70% consensus.
That reading is approximate. Economists Wolfers and Zitzewitz found that prices are usually close to traders' average belief, typically within a percentage point, with the largest gaps near $0 and $1. The number shown on a market page may also not be a price you can trade at: on Polymarket a 34¢ bid and a 40¢ ask show as 37%, while buying costs at least 40¢ plus any fee. In the Kalshi example above, the trade only has a positive expected value if the true chance is above $41.68 ÷ $100.00 = 41.68%. See prices are probabilities: how to read prediction market odds for conversions to betting odds and more worked examples.
Versus polls. A poll asks a sample of people what they intend or prefer; a prediction market asks traders to put money on what they expect to happen, and the price pools what they know. Wolfers and Zitzewitz describe markets as able to "aggregate disperse information into efficient forecasts". A poll measures opinion at one moment; a market forecasts an outcome, and it can be wrong too.
Versus stocks. A share of stock is an ownership stake in a company with no expiry date. An event contract is a derivative that expires on a set date and pays a fixed amount, $1 or $0, depending only on whether the event happened. In the US, event contracts are overseen by the CFTC under commodities and derivatives law, not by the securities regulator.
Versus sportsbooks. A bookmaker sets the odds itself, takes the other side of your bet and builds its margin into the odds. An exchange instead matches customers against each other and charges a fee. A Harvard Business School study of the UK market found that Betfair's exchange margin (the over-round) was typically under 3%, against about 12% at bookmakers. Prediction-market exchanges follow the exchange model, which is why the UK regulator compares them to betting exchanges (see below). Critics dispute Kalshi's framing that it isn't acting as the house, pointing to affiliated market makers and fees, and argue the product works like sports betting.
How they compare:
| Type | Your counterparty | Who sets the price | How you win or lose | Who regulates it (examples) |
|---|---|---|---|---|
| Prediction market (US-regulated exchange) | Another trader on the opposite side; the exchange only matches you | Traders, through bids and offers on the order book | Each contract pays $1 if your side is right and $0 if not; fees on top | The CFTC at federal level; many states dispute sports contracts |
| Opinion poll | None: nothing is bought or sold | No price: the pollster reports the share giving each answer | Nothing is staked | Not a traded financial product |
| Stock | Another investor, matched by the exchange | Buyers and sellers | Value rises or falls with the share price; no fixed payout or end date | Securities regulators (the SEC in the US), not the CFTC |
| Sportsbook (bookmaker) | The bookmaker itself | The bookmaker, with its margin built into the odds | Win the payout at the odds you took, or lose your stake | US: state gaming regulators, licensed state by state |
| Betting exchange (e.g. Betfair) | Another bettor: one backs, the other lays | Bettors' offers | Win at the matched odds or lose your stake; commission on winnings | UK: the Gambling Commission |
There is no single answer: it depends on the regulator and the country. PolyKal reports how each one classifies them and doesn't take a side.
US federal level. The CFTC treats event contracts as derivatives, "typically structured as swaps", traded on exchanges it designates (designated contract markets). Exchanges list new contracts by certifying them themselves, and the CFTC can bar contracts involving terrorism, assassination, war, gaming or unlawful activity if it finds them contrary to the public interest.
The 2026 CFTC proposal. A 2024 CFTC proposal to ban political and some sports contracts as "gaming" was withdrawn under Chairman Michael Selig, announced in late January and formally withdrawn in February 2026. On June 10, 2026 the CFTC proposed a new framework: it defines "gaming", treats elections and awards as contests rather than games, presumes that contracts on scores, spreads and statistics in professional and college sports are allowed, and presumes that contracts on injuries, officiating, "first action" events and pre-college sports are not. The rule is a proposal; public comments closed on July 27, 2026. The law firm DLA Piper reports that 44 states jointly disputed the CFTC's authority in their comments.
US states. Many states say sports event contracts are sports betting that needs a state license, and have sent cease-and-desist letters or gone to court. The courts disagree. In April 2026 the Third Circuit sided with Kalshi in a New Jersey case, finding state gaming law likely preempted; on August 28, 2026 the Ninth Circuit ruled 3–0 for Nevada that sports event contracts are not "swaps" and that Nevada law is likely not preempted. DLA Piper counts 19 states in proceedings, and the CFTC itself has sued several states over the question. In early September 2026 New Jersey asked the Supreme Court to decide, as did Crypto.com and Robinhood after losses in the Ninth Circuit; we found no decision to hear the case as of September 24, 2026.
Outside the US. Regulators elsewhere generally treat prediction markets as gambling. The UK Gambling Commission wrote in February 2026 that their "core aspects are akin to what in the UK would be described as a 'Betting Exchange'", so operators need a gambling license to serve Great Britain, while financial products fall to the Financial Conduct Authority. Nine European gambling regulators issued a joint statement in June 2026, and authorities in countries including France, Italy, Spain, Brazil and India have ordered blocks of Polymarket, Kalshi or both.
As of September 2026 the main platforms fall into a few groups: exchanges regulated by the CFTC in the US, brokers and apps that pass customers' orders to those exchanges, an international crypto-based platform, and play-money or academic markets. The table lists only platforms our research covers.
| Platform | Type | Who can use it | Money |
|---|---|---|---|
| Kalshi | US exchange regulated by the CFTC since 2020 | US users (several states dispute its sports contracts) and, since October 2025, users in many other countries; its member agreement lists 55 restricted countries | US dollars; international members can also deposit by debit card, wire or crypto |
| Polymarket (polymarket.com) | International crypto-based exchange on the Polygon blockchain; users hold their own wallets | Adults (18+) outside the US; its help center lists dozens of blocked countries, including the US, the UK and France | Crypto: pUSD, a token backed 1:1 by USDC |
| Polymarket US | Separate US exchange regulated by the CFTC; accounts aren't connected to polymarket.com | US residents after an identity check, in the iPhone app (an Android version is listed on Google Play, though Polymarket's help center still says it is coming soon); its App Store listing says trading is unavailable in Nevada | US dollars (debit card, bank transfer or wire) |
| Robinhood | Broker (futures commission merchant) that routes orders to exchanges, including Kalshi and ForecastEx | Robinhood's US customers | US dollars |
| ForecastEx (Interactive Brokers) | CFTC-registered exchange and clearinghouse owned by Interactive Brokers, launched in August 2024 | Interactive Brokers clients through its ForecastTrader platform; also reachable through Robinhood | US dollars |
| FanDuel Predicts (CME Group) | App offering contracts listed by CME Group, launched December 22, 2025 | Launched in five states (AL, AK, SC, ND, SD) with a wider rollout planned; sports contracts only where online sports betting isn't legal | US dollars |
| Manifold | Play-money market | New users get Ṁ1,000 of play money; anyone can create a market | Play money ("Mana") that can't be converted to cash |
Others include Crypto.com's CFTC-registered exchange (sports contracts since December 2024), Rothera, a Robinhood–Susquehanna venture that listed its first contracts in May 2026, PredictIt and the Iowa Electronic Markets. In the UK, Betfair runs a betting exchange under a Gambling Commission license. For the two biggest side by side, see Polymarket vs Kalshi compared, our Polymarket review and our Kalshi review.
Trading has grown fast, mostly in sports. Monthly volume on Kalshi and Polymarket combined, as counted by The Block and reported by the Pew Research Center, went from under $5 billion in September 2025 to about $24 billion in April 2026 and a peak of $53.0 billion in July 2026, during the World Cup: $40.1 billion on Kalshi and $12.89 billion on Polymarket and Polymarket US. In August 2026 it fell 14.5% to $45.33 billion, the first monthly drop in a year (The Block, September 2, 2026).
Counting methods differ. The Block and Pew count notional volume on the taker side only, valuing each contract at $1. TRM Labs counts both sides of each trade and reported more than $20 billion a month in January 2026, against about $1.2 billion a month in early 2025. Kalshi's own "annualized" figures (a rise from $52 billion to $178 billion in six months, self-published in May 2026) are extrapolations, and Robinhood's self-published figures count contracts (13.6 billion in Q2 2026), not dollars. A Columbia University study estimated that about 25% of Polymarket's historical volume was likely wash trading, meaning trades with no real change of position. These figures can't be compared directly.
Where the money goes. From July 2024 to April 2026, 80% of Kalshi's volume was sports and 4% politics; on Polymarket's international site it was 39% sports and 32% politics (Pew). Pew found that prediction-market sports volume in the first quarter of 2026 was about on par with the amount wagered at US legal sportsbooks. Investors valued Kalshi at $22 billion in May 2026 and Polymarket at about $21 billion in a round reported at the end of August 2026.
Across many markets, often quite accurate, but not always. Compared with 964 polls in five US presidential elections from 1988 to 2004, the Iowa Electronic Markets were closer to the final result 74% of the time (Berg, Nelson and Rietz, 2008), and a 2026 study of Kalshi found prices broadly match how often events happen, getting more accurate as expiry nears. But a working paper on more than 2,500 markets in the 2024 US election found identical contracts priced differently across exchanges, and the platforms' own accuracy pages don't disclose their full method. See how to read prediction market odds for the evidence in detail.
These problems are documented; the sources are listed at the end of the page.
It depends on your country and, in the US, your state, and the rules change often. Kalshi's member agreement lists 55 restricted countries, polymarket.com blocks the US and dozens of other countries, and governments have ordered blocks that go beyond both platforms' own lists. See where Polymarket and Kalshi are available for the current, dated list.
Polymarket's Terms of Use prohibit using a VPN or similar tool to get around these restrictions, and doing so can also break local law. PolyKal doesn't help anyone get around them.
A prediction market is an exchange where people buy and sell contracts on future events, such as elections, economic data or sports results. A typical contract pays $1 if the event happens and $0 if it doesn't, so its price between $0 and $1 is read as the market's estimate of the probability: 70¢ means about 70%.
You buy Yes or No contracts from other traders through the exchange's order book, at prices set by supply and demand. Before the end you can usually sell at the current bid; at the end, the market resolves under its written rules, and each winning contract pays $1 while each losing one pays $0. The exchange usually charges fees and, on US-regulated venues, doesn't take a side.
It depends on where you are, and this isn't legal advice. In the US, Kalshi and Polymarket US are exchanges regulated by the CFTC, but many states argue their sports contracts are unlicensed betting, federal appeals courts have split, and Supreme Court petitions were filed in September 2026. Many other countries treat prediction markets as gambling and have blocked some sites. See where Polymarket and Kalshi are available.
Regulators disagree. The US CFTC treats event contracts as derivatives and in June 2026 proposed rules that treat elections as contests rather than games; the Ninth Circuit found in August 2026 that sports event contracts are not "swaps" and that Nevada's gaming law is likely not preempted; and the UK Gambling Commission says prediction markets are akin to betting exchanges. Whatever the label, you can lose the full amount you put in.
Often fairly accurate across many markets, especially close to the end: the Iowa Electronic Markets were closer to the result than polls 74% of the time across five presidential elections, and prices on Kalshi broadly match outcomes. But cheap longshots win less often than their price suggests, the same event can trade at different prices on different sites, and a single result can't prove a market right or wrong. See how to read prediction market odds.
Kalshi is a US exchange regulated by the CFTC since 2020; it uses US dollars and also accepts users in many other countries. Polymarket's main site, polymarket.com, is a crypto-based platform using pUSD, a USDC-backed token, and doesn't accept US residents, who use the separate, CFTC-regulated Polymarket US app in dollars. Fees differ too: Kalshi's taker fee is 0.07 × contracts × price × (1 − price), and polymarket.com's rate runs from 0 to 0.07 depending on the category. In August 2026 Kalshi traded $37.17 billion, against $8.16 billion for Polymarket and Polymarket US (The Block). See Polymarket vs Kalshi compared.
Some traders do, but on average traders lose. Prices already reflect what the crowd knows, so you only gain if you judge the odds better than everyone else; fees and bid-ask spreads cost money on every trade; and many cheap, small-probability contracts lose. A study of Kalshi found an average return of about −20% on contracts and about −31% for takers, and a working paper on Polymarket found that fewer than one-third of traders make a profit. PolyKal doesn't give trading advice.
Traders do. The price is where buyers and sellers agree to trade on the order book, so it moves with every trade; the exchange lists the market, matches orders and settles the result. That is the main difference from a sportsbook, where the bookmaker sets the odds and takes the other side.
Not on a standard $1/$0 contract that you buy: the most you can lose is what you paid plus fees, because the contract can't be worth less than $0. Polymarket and Crypto.com, for example, describe their contracts as fully backed by collateral. You can, however, lose all of it.
By trading volume, Kalshi was well ahead of Polymarket in July and August 2026: in August The Block counted $37.17 billion on Kalshi and $8.16 billion on Polymarket and Polymarket US. In March 2026 Pew found the two roughly level, with about half of total volume each. Rankings depend on how volume is counted.
Every fact on this page comes from the sources below, checked on September 24, 2026. Rules, court cases, fees and volumes change often, so figures are as of September 2026, and we re-verify this page quarterly. Statistics that platforms publish about themselves are labeled as self-published, and where sources disagree we say so or leave the point out. See our methodology.
Regulators and courts: CFTC: Understanding Prediction Markets and Event Contracts, CFTC: Contracts & Products, CFTC: Kalshi designated as a contract market (2020), CFTC: order against Polymarket (2022), CFTC: complaint against Intrade (2012), CFTC Letter 25-20 (PredictIt, 2025), Federal Register: Prediction Markets, advance notice (March 2026), CFTC: proposed rule (June 2026), Federal Register: proposed rule (June 2026), D.D.C. decision in KalshiEX v. CFTC (2024), D.C. Circuit order in KalshiEX v. CFTC (2024), Third Circuit opinion in KalshiEX v. Flaherty (2026), Ninth Circuit opinion in KalshiEX v. Assad (2026), US Justice Department charge (April 2026), UK Gambling Commission: prediction markets (February 2026), German gambling regulator (GGL): joint statement of European regulators (June 2026), ANJ (France): Polymarket blocking order (July 2026), DGOJ (Spain): proceedings against Polymarket and Kalshi (May 2026), Iowa Electronic Markets FAQ.
Platforms: Kalshi: What Is Kalshi?, Kalshi help: How are prices determined?, Kalshi fee schedule, Kalshi member agreement, Kalshi help: trading from outside the US, Kalshi: $1 billion raise (self-published), Polymarket help: What is Polymarket?, Polymarket help: How Are Prices Calculated?, Polymarket help: How Are Markets Clarified?, Polymarket help: Can I Sell Early?, Polymarket help: Does Polymarket Have Trading Limits?, Polymarket help: Geographic Restrictions, Polymarket help: Website vs App, Polymarket help: Is My Money Safe?, Polymarket docs: prices and order book, Polymarket docs: fees, Polymarket docs: resolution, Polymarket docs: how positions work, Polymarket docs: pUSD, Polymarket Terms of Use, Polymarket US docs: what is Polymarket US, Polymarket US docs: deposit methods, App Store: Polymarket, Polymarket: market-integrity rules (March 2026), Robinhood newsroom (September 2026), Robinhood: prediction markets hub (March 2025), ForecastEx launch release, CME Group: FanDuel Predicts, Crypto.com help: sports event trading, Crypto.com: sports launch (December 2024), Manifold: About, Betfair: What is a betting exchange?.
Research: Wolfers & Zitzewitz, "Prediction Markets" (2004), Wolfers & Zitzewitz, "Interpreting Prediction Market Prices as Probabilities" (2006), Arrow et al., "The Promise of Prediction Markets" (2008), Berg, Nelson & Rietz, "Prediction market accuracy in the long run" (2008), Bürgi, Deng & Whelan on Kalshi (VoxEU, 2026), Clinton & Huang on the 2024 election markets (working paper), Reichenbach & Walther on Polymarket (working paper), Rothschild & Sethi on Intrade, Tsang & Yang on Polymarket's 2024 market (preprint), Casadesus-Masanell & Campbell on Betfair (Harvard Business School), Rose-Berman, "Kalshi's Favorite Lie" (opinion).
Data and reporting: Pew Research Center (May 2026), Pew Research Center (September 2026), The Block: August 2026 volume, TRM Labs on monthly volume, DLA Piper tracker (September 2026), Ropes & Gray on the CFTC proposal, Corporate Compliance Insights on the withdrawn 2024 proposal, Insider Sport on the Supreme Court petitions, Decrypt on the 2024 election market, Decrypt on the Columbia wash-trading study, NBC News on the French trader, NPR on Kalshi's insider cases, CoinDesk on the Ukraine minerals market, Decrypt on the Zelenskyy suit market, CasinoBeats on the Khamenei market, CoinDesk on the June 2026 security incident, TechCrunch on Kalshi's valuation, Bloomberg on Polymarket's valuation, PBS NewsHour on the CFTC suing states (April 2026), Lefosse on Brazil's rules (April 2026), ThePrint on India's blocking orders (May 2026), Punto Informatico on Italy's blacklist (July 2026).
History: Quartz on Intrade's shutdown (2013), CCH summary of the PredictIt ruling (2025), Wikipedia: PredictIt (secondary), Yahoo Finance on the CFTC dropping its Kalshi appeal, PR Newswire: Polymarket acquires QCEX, CoinDesk on Kalshi's expansion to 140 countries, CoinDesk on the Polymarket US app launch, Yahoo Finance on Polymarket US opening to all iPhone users, DeFi Rate on Rothera (secondary), ICE announcement naming Polymarket's founding.
PolyKal is an independent research tool, not affiliated with Polymarket, Kalshi or any platform named here. This page explains what prediction markets are; it is not financial, legal or tax advice, and not an invitation to open an account or trade.