A Los Angeles restaurant group turned to the prediction market Kalshi to protect itself against a disappointing Dodgers postseason, trading $60,000 on the outcome of the team’s playoff run. The hospitality company, which operates 25 bars including several near Dodger Stadium, placed the position on whether the Dodgers would fail to reach the World Series. When the team lost to the Milwaukee Brewers in the National League Championship Series last week, the trade paid out nearly twice the original stake.
The payout helped offset revenue the business would have lost had the Dodgers played more home playoff games. Kalshi confirmed the trade to Fortune but did not name the restaurant group. The episode is one of a growing number of examples in which companies are using prediction markets not as sports betting venues but as financial hedges against specific business risks.
Joshua Mitts, a Columbia University law professor who studies prediction markets and securities law, said the restaurant group’s trade was «functionally quite similar to a form of bespoke insurance.» He added that other businesses facing comparable risks might participate in similar markets if suitable contracts were available.
Prediction markets have surged in popularity this year. Trading volume on the platforms doubled between May and July, according to Pew Research Center data, topping $58 billion on Kalshi and nearly $22 billion on Polymarket, driven primarily by sports. Part of the appeal is looser regulation: the sites are classified as financial exchanges rather than gambling platforms, which allows them fewer restrictions than online sportsbooks.
Small businesses searching for unconventional risk-management tools appear to be among the growing users. In June, a New York City bar called The Jeffrey promised patrons it would cover every tab up to $100 if the Knicks won an NBA Finals game. The bar placed a $5,000 bet on Kalshi that the underdog Knicks would win, and a $13,500 payout allowed it to pay off customers’ tabs and still net $8,500. Last month, an unnamed third-party insurance company placed five trades on Kalshi that could pay out $3 million if Louisiana State University wins this season’s national championship, matching what the college’s football program would owe coach Lane Kiffin in bonuses. Earlier this summer, a Los Angeles ice cream shop reported making nearly half its monthly rent, up to $1,500, by betting on when cold fronts would arrive, hedging against slow demand.
Platforms have begun adapting to these uses. In August, independent financial economist Lauris Zminsky launched an app called Blanket in partnership with Kalshi, an AI engine that suggests specific yes-or-no markets that could offset a risk for a particular business. Zminsky described it as a reasoning and discovery tool that funnels users toward Kalshi, and said he wanted to build a self-service option for small-business owners.
Prediction markets lend themselves to risk management because they incorporate large samples of data, especially for sports-related hedging, said Eric Zitzewitz, an economics professor at Dartmouth College. «There’s so much data, there’s so much repetition, there’s so many games all the time that they can be very well-calibrated,» he said. Users pay fees for the insurance the markets provide, but Zitzewitz said they should be getting fair underlying prices.
Mitts noted that the approach works best for yes-or-no outcomes such as weather or a game, rather than hedges based on specialized knowledge, where a competitor with better information could make a business owner a «systematic loser.» The tool is also likely to work best in major metropolitan areas with large trading markets, while still serving idiosyncratic local purposes.
Zitzewitz said the mass sampling and real-time information gathering on prediction markets can also inform broader business decisions. An e-commerce retailer, for example, could look at trades tied to future Amazon credit card receipts for a fresher forecast of consumer spending than GDP growth provides.
Still, Mitts argued that the platforms’ light regulation may prevent them from becoming reliable insurance substitutes. Insurers have extensive resources to prevent fraud, while the decentralized structure of prediction markets and their evolving regulatory environment leave them without the same guardrails. «This is a bit of a free-for-all,» he said.