With so-called "prediction markets" giving taxpayers new ways to place wagers on future events such as sports, elections, pop culture happenings and even the weather, the uncertainty isn't limited to the predictions — users are taking tax risks as well.
Under the One Big Beautiful Bill Act, gambling losses are only deductible up to 90% of gambling winnings. For example, if a taxpayer has $10,000 of gains and $10,000 of losses, they still must report $1,000 of income.
Prediction markets claim they are not gambling platforms, but instead are financial contracts regulated by the Commodity Futures Trading Commission, not by the states. Under this analysis, losses for taxpayers who use CFTC-approved platforms such as Kalshi and Polymarket may be taxed differently than a taxpayer who made the same prediction on the same event at a casino, and are subject to gambling loss limitations.
The Internal Revenue Service has been silent in this area, but is not bound by either the CFTC, which tends to side with the position that prediction markets are financial contracts, nor the position of some states that have sued platforms offering prediction markets, claiming they are gambling. This silence is forcing taxpayers to decide on their own how they want to treat the gains and losses.
"It is hard to tell where the line is," said James Creech, a principal with Baker Tilly's specialty tax practice. "If I make a prediction on who will win the Super Bowl on an app rather than at a casino, the tax treatment could be completely different. Even if I am making lots of small bets, the differences in the tax treatment can have meaningful impact over time."
"Digital providers like Polymarket and Kalshi integrate with apps that you can get on your phone like Robinhood," he explained. "They have really broad reach. With these, I can wager on things like who will win a sporting event, or what the weather will be. I can place a wager or a money prediction on what the high temperature will be in Phoenix tomorrow. And because these are pushed through financial products, they argue that they are governed by the Commodity Futures Trading Commission, and that states don't have the ability to regulate them. They're also available through other service providers that operate as brokerage accounts — they're likely to issue 1099s or 1099-Bs for these prediction markets, where if the bet was made for the same exact dollar kind of prediction about a future event at a casino, they would get a W-2G. and now, with the [One Big Beautiful Bill Act], the gambling losses for the wager at the casino are going to be perceived to be different than the wagering losses on the app."
"There's kind of a mismatch between the expectations of what I can do on an app should be versus what I can do in the casino, and how the tax consequences can be different," he added.
Common sense might suggest that they should be treated the same, but of course, that's not always the case with the federal income tax.
"One of the things that I don't understand is why the IRS has been silent on this," Creech said. "If you're betting that the Packers will beat the Lions, it's the same intuitive mental exercise, and just because one of them calls itself a financial product or a derivative, it's still the same event that the derivative is based on. Since the IRS in not bound by the CFTC, it could say that anything to do with a sporting event, or anything that does not have an element of skill is going to be gambling and it doesn't matter if it is reported on a W-2G or a 1099-B. It could come as surprise to a lot of people who win and lose small amounts frequently. But when you start adding up the aggregate total of what they gamble, the numbers get quite large."
Ultimately, the IRS is going to address the issue. Meanwhile, Creech suggests that until the agency gives guidance, practitioners should follow the 1099.
"Otherwise you have to impute 10% of income and will probably alienate clients," he said. "There is not an easy answer here. Everybody's going to have to come up with an answer depending, in part, on how risk-adverse they want to be."
