Wall Street has quietly taken over the business of guessing who wins elections, and the change is moving faster than most voters notice. Kalshi, Polymarket and a handful of newer prediction markets have spent the past eighteen months turning political outcomes into tradable contracts, signing data deals with CNN, CNBC, Fox, Dow Jones and the Associated Press while pulling billions in volume from anyone with a brokerage account or a crypto wallet.

The reason is simple: after the polling industry stumbled through 2016, 2020 and 2024, money looked like a more honest signal than telephone surveys, and a regulatory shift in late 2024 finally let Americans bet on it legally.

That last part matters more than the marketing copy admits, because a federal judge ruled in September 2024 that Kalshi’s election contracts were commodity derivatives rather than gambling, and the CFTC followed by approving Polymarket as a regulated US exchange in November 2025. Suddenly, the same Commission that polices soybean futures was overseeing wagers on who runs the Senate.

The cast goes beyond two startups: Intercontinental Exchange, owner of the New York Stock Exchange, committed up to $2 billion to Polymarket in October, while Kalshi closed a $1 billion round at an $11 billion valuation around the same time. Both companies now sit on cap tables crowded with the same venture firms that financed sports betting, and Donald Trump Jr. has joined both as an adviser. The category is no longer fringe.

From the trading floor to prime time

The tipping point arrived in late 2024, on contracts tied to the Trump–Harris race, when Polymarket alone handled north of $3.5 billion in volume on that single election and its odds called the winner days before any pollster was willing to. Newsrooms noticed, and by December 2025 the partnership deals were tumbling out one after another, each promising real-time tickers and on-air graphics built from prediction market prices.

Where the data shows up is the part that should worry editors more than it apparently does. CNBC now runs a Kalshi ticker alongside its programming, CNN bakes its probabilities into political segments, and The Wall Street Journal cited Polymarket’s odds in a March story about whether US forces would enter Iran; within days, its parent company signed an exclusive deal with the platform. Fox News added Kalshi in April. TheWrap traced the cascade deal by deal, and the pattern is consistent: the odds appear inside the news, not beside it.

Why this matters comes down to incentives, since a poll measures what people say they want while a market measures what people with capital think will happen, which is a different question and was never meant to be the same one. For the curious reader, comparative breakdowns of the leading prediction markets lay out how each one is structured, regulated and priced, useful context because the resulting probabilities are now being broadcast back to voters as if they were public opinion.

The wisdom of the crowd, audited

The trouble is that the authority is mostly aesthetic. Recent Vanderbilt research analysing 2,500 political markets across Kalshi, Polymarket, PredictIt and the Iowa Electronic Markets, covering more than $2 billion in trades during the final weeks of the 2024 race, found that Polymarket got only 67% of its markets right, Kalshi 78%, and the academic PredictIt 93%. Identical contracts diverged in price across exchanges, and arbitrage gaps widened in the final fortnight. That is not the behaviour of an efficient information machine; it is the behaviour of a thin, hype-driven derivatives venue.

Single traders can also move the price, and one pseudonymous account held more than 20% of all “yes” shares on Trump in 2024. As this outlet noted at the time, the contracts swung sharply toward Trump in the final weeks even as national polling averages still gave Harris a slight lead. The market was telling a different story, and the market turned out to be right, though critics argue it was right by accident.

Insider activity adds another layer: NPR reported this month that campaign staffers routinely bet on their own candidates using private polling, calling the practice “commonplace”, and Kalshi has already suspended candidates caught wagering on themselves. An Army Special Forces soldier involved in the Maduro raid was arrested in late April for using classified information to pocket more than $400,000 on Polymarket, and hours before US strikes on Iran, anonymous traders piled $1,000-plus bets onto the exact outcome. The exchanges call this insider trading and promise to police it, though the evidence so far suggests they cannot.

The shift is also financial in the literal sense, because Kalshi’s chief executive has said the company’s goal is to “financialize everything and create a tradable asset out of any difference in opinion.” Worth reading twice. It treats public life, from wars and pardons to elections and cabinet shake-ups, as a class of underlying assets, and the news outlets selling viewers on the resulting odds are now commercial partners of the people writing the contracts.

What gets lost is the difference between predicting an election and trading one: the first is a research problem, the second a profit centre, and a democracy that confuses them ends up letting the most leveraged participants set the narrative for the rest. If the conventional wisdom on cable news is now whatever a few thousand anonymous accounts with capital decide it should be, who exactly is being polled?