Prediction Markets vs Sportsbooks: 2026 Data Report

Prediction markets went from a niche product to a rival to the US sportsbook industry in roughly a year. Platforms such as Kalshi and Polymarket now list event contracts on nearly every major game, and in the months covered here they have handled more sports-related dollar volume than the whole legal sportsbook sector in some periods. This report pulls together the published numbers on volume, fees, taxes, accuracy and legal fights into one reference page, and flags where the sources disagree.

If you are new to the topic, our betting basics guide explains odds and implied probability, our US sports betting overview covers state-regulated sportsbooks, and our sports betting hub links to every sportsbook comparison we maintain.

Key Findings: Prediction Markets vs Sportsbooks

Key findings: Combined Kalshi and Polymarket volume rose from under $5 billion in September 2025 to about $24 billion in April 2026 and $50.6 billion in July 2026. Sports are about 80% of Kalshi volume. A taker pays roughly $2.50 to $3.50 per $100 on a 50/50 game versus about $4.55 at -110. The CFTC proposed case-by-case review in June 2026, and six states have enacted prediction market laws.

How Fast Have Prediction Markets Grown?

Volume is the headline number, but it is also the most contested one. Trackers count differently: some count the dollars traded, some count contracts at their full $1 face value, and some count only taker volume. A tracker that counts every contract at face value reported $188 billion for July to September 2026, while Macquarie put total prediction market volume at $20.1 billion for September 2026 alone. Both can be true because they measure different things. The table below uses each source’s own definition and names it.

Two comparisons are worth highlighting. In April 2026 the two platforms handled about 1.7 times the roughly $14 billion monthly average of legal US sportsbooks in 2025, and by July the $50.6 billion combined figure was more than three times that average. The caveat is that sportsbook handle counts every bet at its stake, while an exchange volume figure can count the same position being bought and sold several times, so the two are not a like-for-like measure of customer wagering.

Kalshi vs Polymarket

Kalshi overtook Polymarket in October 2025, when it posted $4.39 billion against $3.02 billion. By June 2026 the gap was about three to one, with Kalshi at $31.5 billion and Polymarket at $10.8 billion. Kalshi held about 52.6% of combined 30-day volume in March 2026. Its weekly volume reached $19 billion in the most recent week reported in October 2026, with a daily record above $4 billion on October 11. Polymarket’s weekly volume was $3.7 billion. In April 2026, Polymarket’s US arm handled $1.3 billion against $9 billion on its international platform.

How Much of the Volume Is Sports?

Sports are the engine of the growth, especially on Kalshi.

Polymarket is more diversified: crypto and politics together made up 52% of its turnover, and its largest single markets in mid-2026 included the FIFA World Cup winner (over $3 billion), the NBA champion (over $414 million), the English Premier League winner ($322.6 million) and Federal Reserve rate markets (over $103.7 million). On the opening Sunday of the 2026 NFL season, about $1.8 billion of the $3.3 billion traded was tied to NFL games, according to PredictDesks.

Fees: What Does a Prediction Market Bet Cost?

A sportsbook earns money from the margin built into its odds. A prediction market earns money from a per-trade fee. At standard -110 odds on both sides of a game, the implied probability of each side is 52.4%, the combined overround is 4.8%, and the book’s hold is about 4.5%. The table compares that cost with exchange fees for a $100 bet on a 50/50 game.

Exchange fees are not flat. Kalshi’s formula is 7% x contracts x price x (1 - price), which makes the fee highest as a share of the stake on long shots and lowest on heavy favorites. Polymarket’s sports taker rate is 0.05 in the same formula. As a check, Forbes’s example of 10,000 contracts at 42 cents gives a Kalshi fee of $170.52.

The pattern matters for bettors. Parlay-style combos on Kalshi were priced at an average implied chance of about 9%, compared with over 40% for single markets, which is exactly the zone where the fee share is highest. Our sportsbook promotions page shows what bonus value a regulated book can offset against its margin.

Taxes: Why the Economics Differ

State sportsbook taxes range from 6.75% in Nevada to 51% in New York and Rhode Island, on top of the 21% federal corporate rate. Prediction markets are regulated by the CFTC rather than state gaming commissions, so they have not paid those state rates. Forbes modeled New York as an example.

That $43.75 difference per $100 is the center of the industry fight. The American Gaming Association says prediction market sports contracts have already cost states more than $1 billion in lost sports betting tax revenue. States are responding: Kentucky enacted a 14.25% excise tax on operators’ transaction fees in April 2026, and a North Carolina budget proposal would tax net trading fee revenue at 6%.

Are Prediction Market Prices Accurate?

Because prices are probabilities, calibration is a natural test. A PredictionScout study of 33,716 settled questions from January 1, 2025 to September 30, 2026 compared the favorite’s price one day before the outcome was clear with the actual result.

Favorites won 74.0% of the time overall, and heavy favorites priced at 95 cents or more won 97.3%. Of 12,609 heavy favorites at 90 cents or more, 590 lost, which is 4.7% and consistent with prices that were close to fair. Our implied probability guide explains how to convert any price or odds into a probability for your own comparison.

Regulation and the Courts

The CFTC released a 267-page proposal on June 10, 2026 that would review each event contract individually. It named sports examples likely to be prohibited, including injury occurrence or severity, refereeing decisions, in-game physical altercations and youth sports, and said in-game props describing discrete actions by specific participants would likely be barred. A 45-day comment period followed. The agency usually has five commissioners but at the time had one, Chairman Michael Selig. The leagues have weighed in too: the NBA and NFL have both asked regulators or operators to restrict trades on manipulable events such as officiating, injuries and fan actions.

PredictionScout counts 84 federal lawsuits and 22 appeals involving Kalshi, Polymarket, Crypto.com or Robinhood, 59 of them filed in 2026, with state parties in cases across 20 states. It also tracks 79 state bills in 15 states; six states have enacted one: Connecticut, Illinois, Kentucky, Minnesota, North Carolina and Tennessee. In the Senate, Sens. John Curtis and Adam Schiff introduced a March bill to ban event contracts that resemble sports bets. Because court dates in this area change quickly, check the primary dockets before relying on any date above. Also review whether your state is among the roughly 39 with legal sportsbooks; about 11 states have none, which is where exchanges have an availability advantage.

Structural Differences at a Glance

An SSRS poll in May 2026 found that 5% of US adults use prediction markets for sports, rising to 11% of men aged 18 to 34. On the midterms, traders had wagered over $220 million through September 30, 2026, against $92 million in the 2024 cycle, across more than 21,000 markets versus 463.

Methodology

We compiled figures from published 2026 reporting: ESPN’s coverage of the CFTC proposal, Predictions Market Fans’ volume comparison, Blockonomi’s statistics roundup (which relays Pew, Macquarie, PredictionScout and Forbes numbers), PredictDesks’ guide to sports prediction markets, and SCCG Management’s June volume note. We did not independently audit exchange data. Where trackers disagreed we show both and name each source. The only figures we calculated ourselves are the -110 overround and hold, the fee-by-price table (derived from the published fee formulas), the 87% March Kalshi sports share (9.9 / 11.39), and the 95.3% win rate for 90-cent favorites (derived from 590 losses out of 12,609). Data runs through mid-October 2026.

Limitations: volume definitions differ across trackers, some figures are relayed second-hand, and legal dates are moving targets. This page is informational and is not legal, tax or betting advice.

Frequently Asked Questions

Are prediction markets bigger than sportsbooks?

By reported volume, yes in some periods, but the measures differ. Q1 2026 prediction market sports volume was about $38.6 billion against about $40 billion at legal sportsbooks, while July 2026 sports volume on Kalshi and Polymarket was about $42 billion.

Which is cheaper to bet on, a prediction market or a sportsbook?

On a 50/50 game, a Kalshi market order costs about $3.50 per $100 and a Polymarket market order about $2.50, versus about $4.55 at -110. Limit orders can be cheaper still.

What share of Kalshi volume is sports?

About 80% since July 2024 according to Pew, 87% in March 2026 and 79% in September 2026.

Do prediction markets pay state gambling taxes?

Not at the rates sportsbooks do. The CFTC, not state gaming commissions, regulates them, though Kentucky enacted a 14.25% excise tax on transaction fees in April 2026 and the measure is under court challenge.

Is sports event contract trading legal in every state?

No. Rulings have gone both ways: the Third Circuit sided with Kalshi on April 6, 2026, while the Ninth Circuit ruled against it on August 28 and the Sixth Circuit sided with Ohio and Tennessee on September 25.

Sources

• ESPN: CFTC proposes rules limiting prediction markets
• Predictions Market Fans: Is Kalshi or Polymarket Bigger?
• Blockonomi: Prediction Market Statistics 2026
• PredictDesks: Sports prediction markets explained
• SCCG Management: Kalshi and Polymarket volume hits $45 billion in June

Cite This Article

If you use data from this article, please link back to https://www.deucescracked.com/blog/prediction-markets-vs-sportsbooks-volume-fees-data-2026. More analysis is on our blog.