Prediction markets are booming, but a Wall Street Journal review suggests the gains are flowing to a small elite. Activity hit a record last month, even as most users on Polymarket and Kalshi struggled to stay ahead.
Summary
Volume surges as losses mount
Monthly notional volume in the sector reached $29.8 billion, an all-time high and about 588% above the same month last year. That surge has been driven by rapid user growth on platforms such as Polymarket and Kalshi.
However, the WSJ found that more than 70% of Polymarket users are losing money. The analysis used platform data and reviewed 1.6 million accounts that had been active since November 2022.
Profits remain concentrated
The report says only 0.1% of accounts captured 67% of all profits. In practice, fewer than 2,000 accounts collectively netted nearly $500 million.
For typical users, the picture was far weaker. The median account is down between $1 and $100, while the bottom 10% of traders lost an average of $4,000 each. Moreover, the WSJ described a split between casual traders and more sophisticated participants with access to better data.
In that context, prediction markets have become a clear example of trading profit concentration, where a small group captures most of the upside. The report pointed to trading firms and other advanced players as the main winners.
Research points in the same direction
An academic study covering November 11, 2022 through March 29, 2026 found similar results. It said 68.8% of Polymarket users lost money, while the top 1% accounted for 76.5% of total profits.
That said, Bloomberg reported that more than 100,000 Polymarket accounts have lost at least $1,000 since January 2025. That figure was nearly double the number of wallets with comparable gains.
Kalshi shows a similar pattern
The WSJ also examined Kalshi. A Kalshi spokeswoman said losing users outnumber winners by 2.9 to 1 based on data from the prior month. The company does not publicly disclose total user numbers.
For Kalshi’s mention markets, the WSJ reviewed more than 35,000 completed markets. It found that ‘Yes’ trades priced around a 50% implied winning probability paid out only about 40% of the time.
Moreover, that pattern suggests systematic overpaying by bettors, especially retail users who enter trades at the first price they see. On average, traders placing ‘Yes’ bets at the first displayed price lose about 11% of the money wagered, according to the report.
That result is worse than returns from most Vegas slot machines, based on research from the University of Nevada, Las Vegas. In other words, the latest data shows fast-growing markets, but uneven outcomes for ordinary traders.
