The American midterm elections have always been a high-stakes ritual, but this year the drama is playing out on a new kind of stage. Alongside the television pundits and the endless polling numbers, there are now digital prediction markets—platforms like Polymarket and Kalshi—where anyone with a credit card can wager real money on who will win a Senate seat, control the House, or take a state’s electoral votes. It feels like a game, a futuristic way of engaging with democracy, but it is also something much more serious. The percentages that flash across these platforms look like authoritative data, and they have begun to shape how millions of people understand what is happening in their own country. The trouble, experts warn, is that these numbers are not facts, not certainties, and not official results. They are guesses priced by the crowd, often volatile and vulnerable to manipulation. And in a deeply polarized political environment, where trust in public institutions is already fragile, this blurring of gambling and journalism could be dangerous. The midterms are taking place at a moment when prediction markets are expanding faster than regulators can keep up, and the result is a strange new landscape where the old rules of polling, money, and electoral legitimacy are all colliding at once.
At their core, these prediction markets work like simplified stock exchanges. Users buy and sell contracts whose value is tied to a future event—say, “Republican wins the Arizona governor race” or “Democrats keep the Senate.” If the event happens, the contract pays out; if not, it is worthless. The price of the contract, expressed as a number between zero and one, is interpreted as the market’s implied probability. So when a share in a particular outcome trades at 65 cents, the platform displays a 65% chance. This is clever, intuitive, and, for many, seductively precise. Major media outlets have already begun quoting these numbers alongside traditional polls, treating them as a kind of real-time wisdom of the crowd. Yet the public remains wary. Research has shown that most Americans do not want betting on elections at all—it feels wrong, like turning civic duty into a casino game. That discomfort has prompted action in some places. In Brazil, for example, the National Monetary Council stepped in before the midterms and banned contracts tied to election results and other non-economic events. The rule, which took effect in May 2026, led to Polymarket and Kalshi being blocked in the country. The message was clear: some things are too sacred, or too risky, to be treated as just another financial instrument. But in the United States, the debate is far from settled, and these platforms continue to operate in a legal gray zone, feeding their probability numbers into the national conversation with little oversight.
The deepest worry among experts is not that people will lose money, but that they will lose the ability to tell the difference between a prediction and a promise. Davina Hurt, director of government ethics at the Markkula Center for Applied Ethics at Santa Clara University, has watched this evolution with concern. She argues that prediction markets have already become central to how Americans understand elections, and that this is a dangerous shift. “We are reaching a point,” she says, “where the expectations of the financial markets carry more weight in the public mind than the votes themselves.” That is a stunning statement, but it reflects a real phenomenon. When a platform says a candidate has a 90% chance of winning, that number feels almost like a fact—not a snapshot of wagers but a verdict. People begin to treat it as a guaranteed outcome. And when the actual result differs from the market’s forecast, the temptation is to assume something went wrong. Hurt emphasizes that a high probability is never a promise. “When a prediction turns out to be wrong, it doesn’t mean the choice was wrong,” she says. “However, people will interpret it that way.” This is the heart of the danger. The gap between market expectations and election outcomes can be spun as evidence of fraud, even when no fraud exists. In a country already divided by conspiracy theories about stolen elections, prediction markets are pouring gasoline on the fire, not because they are malicious, but because their numbers are so easily misunderstood.
That is not a hypothetical fear. A vivid example emerged during the Los Angeles mayoral primary in June, when right-wing influencers seized on prediction market estimates to claim the election had been stolen. Republican candidate Spencer Pratt lost badly, but the markets had suggested at some point that his chances of advancing were far better than zero. Those probability estimates were treated as proof that something underhanded had occurred, and the accusations spread quickly across social media. Some of the posts, falsely suggesting electoral fraud, were even labeled as “paid partnerships,” meaning they had been commercially sponsored to reach more eyes. For local election officials, it was a nightmare. Michael Sánchez, spokesperson for the Los Angeles County Election Registration Office, told reporters that the primary had shown exactly how “expectations about an election can be confused with the results.” The episode was a warning: prediction markets do not just record events, they create expectations. And when those expectations are not met, the disappointment can turn into distrust, then into accusations, then into an existential threat to the legitimacy of democracy itself. The Los Angeles case was not a one-off anomaly; it was a preview of what could happen on a national scale, especially in a tight midterm race where control of Congress is at stake and every contested seat will be scrutinized under a microscope.
Even Kalshi, one of the platforms at the center of this storm, seems to recognize the dangers. The company has long maintained that its probabilities are not guarantees and that any attempts to manipulate the numbers are usually offset by the trading of other users. Elisabeth Diana, Kalshi’s communications director, described the platform as “a complement to traditional polling,” not a replacement for it. But actions speak louder than press releases. In a document obtained exclusively by AFP, Kalshi instructed its partners not to question official election results without solid evidence. The company said its partners “should not question the integrity or accuracy of any official vote count result, judicial decision, or other official or formal determination related to the electoral process.” More strikingly, Kalshi warned that it would terminate contracts with any partner who made unsubstantiated claims that an election result was manipulated, rigged, or compromised. This is a remarkable concession—an admission that the platform’s own ecosystem can become a breeding ground for election misinformation. By trying to discipline its partners, Kalshi is acknowledging that prediction markets have real-world consequences far beyond the ledger of profits and losses. Polymarket, the other major player, declined to comment. But the message is clear: platforms that profit from political uncertainty are now trying to distance themselves from the chaos they help create.
The legal and regulatory landscape in the United States remains tangled. Prediction markets are caught between two very different worlds. On one side, they look like commodity markets, where participants trade contracts and take on risk, and where the Commodity Futures Trading Commission might have jurisdiction. On the other side, they look like gambling, and should perhaps be governed by state betting laws, with all the limits and consumer protections that come with casinos and sportsbooks. So far, the courts have not resolved the question, and the platforms operate in a kind of limbo. Some state and local officials are not waiting for a final decision. Jared DeMarinis, who administers elections in Maryland, has asked for a formal evaluation of the legality of election betting in his state. He has also taken the personal step of prohibiting poll workers from betting on election results. His reasoning is simple and powerful: “Now they’ve simply added fuel to the fire by involving money in this. The last thing we want is for the legitimacy of the result to be called into question.” That fear should not be dismissed. In a healthy democracy, elections are the ultimate expression of the people’s will. They are messy, imperfect, and sometimes surprising, but they are supposed to carry an authority that no market can match. The rise of prediction markets does not have to be the end of that ideal, but it does force a reckoning. Every dollar wagered on a political outcome is a small wager on the public trust. And when the numbers are misunderstood, when they are weaponized by partisans, when they shape the news cycle more than the actual votes do, then the entire democratic process is diminished. The midterms will come and go, winners and losers will be declared, and the markets will settle their contracts. But the deeper question remains: in a society obsessed with probabilities, can we still honor the simple, sacred fact of a vote counted and a result accepted? Right now, no one knows.