
Here’s a number that should make every legacy data provider uncomfortable: $21.5 billion. That’s how much money flowed through Polymarket in 2025 alone—up from roughly nothing two years prior. Ninety-five million transactions. Six hundred thousand active users. And perhaps most tellingly, the parent company of the New York Stock Exchange just wrote a $2 billion check to buy in, valuing a prediction market platform at $8 billion.
Let that sit for a second. A decentralized crypto platform where anyone with an internet connection can bet on whether the Fed will cut rates next month is now worth more than most of the polling firms and proprietary data shops that have monopolized this kind of information for decades. The gatekeepers are watching the gates come off the hinges.
For the better part of a century, if you wanted to know what was going to happen—who was going to win the election, whether the economy was heading into recession, how a policy decision would land—you had two options. You could pay a polling firm six figures to survey a few thousand people over landlines, or you could shell out $24,000 a year for a Bloomberg Terminal and hope the analysts were smarter than the market.
Neither option was available to the average person. And increasingly, neither option is very good. Polling response rates have cratered from 36% in the late 1990s to roughly 5% today, according to Pew Research. When 95 out of 100 people hang up on you, the 5 who stay on the line are not a representative sample of anything. The margin-of-error figures that media outlets parrot don’t even account for this non-response bias—they only measure sampling imprecision, which is a polite way of saying they’re measuring the wrong thing entirely.

Polls got Trump wrong in 2016. They got him wrong again in 2020. And in 2024—after pollsters introduced new weighting methods specifically designed to fix their Trump problem—they still called the race a coin flip. The models said toss-up. The bettors said otherwise.
In October 2024, with $3.7 billion wagered on the presidential election, Polymarket had Trump at a 62% probability of winning. The polling averages had Harris at 49.5% and Trump at 48.5%—a dead heat. Polymarket’s odds on election night were 6 to 12 hours ahead of every major news network in calling the race. By midnight, when CNN was still hedging, Polymarket had Trump at 95%.

A 2025 study out of arXiv confirmed what the results already showed: Polymarket was systematically more accurate than polling in predicting the 2024 presidential outcome, particularly in swing states. And this wasn’t a fluke—peer-reviewed research from 2018 and 2020 found that PredictIt consistently outperformed FiveThirtyEight across presidential, Senate, House, and gubernatorial races.
The logic isn’t complicated. When someone answers a poll, they’re telling you who they want to win. When someone places a bet, they’re telling you who they think will win—and they’re putting money on it. Financial incentives have a funny way of clarifying people’s thinking.
Here’s where this gets interesting for us as market participants and students of finance.
The traditional information supply chain works like this: proprietary data firms—your Bloombergs, your Refinitivs, your Morning Consults—collect data, run it through models, and sell the output to institutional clients at institutional prices. A Bloomberg Terminal runs $24,000 per year. A custom polling engagement from a top firm can cost six figures. This information asymmetry has been one of the foundational advantages that institutional investors have held over retail participants for decades.
Polymarket flips this on its head. It aggregates the collective intelligence of hundreds of thousands of participants—each with their own information, their own models, their own edge—into a single, real-time probability. And it’s free. Anyone with an internet connection can see, right now, what the market thinks the probability of a Fed rate cut in March is, or whether the U.S. will enter a recession this year, or who’s going to win the midterms. No terminal. No subscription. No six-figure consulting engagement.

The institutional world is paying attention. In January 2026, Dow Jones—the publisher of The Wall Street Journal, Barron’s, and MarketWatch—signed an exclusive deal to integrate Polymarket data directly into their platforms. Prediction market probabilities will sit alongside stock tickers and earnings calendars. Almar Latour, CEO of Dow Jones, called it “a rapidly growing source of real-time insight into collective beliefs about future events.”
Read that again. The publisher of the Wall Street Journal is embedding a crypto prediction market’s data into its flagship products as a primary information source. That’s not a partnership between equals. That’s an acknowledgment that the crowd’s signal is now competitive with—and in some cases superior to—the institutional apparatus.
Who loses? The uncomfortable answer is: everyone who sells forecasts for a living. Polling firms that charge clients to survey 1,500 people over two weeks are now competing with a platform that aggregates the financial convictions of 600,000 participants in real time. Bloomberg’s terminal still has enormous value for market data, execution, and communication—but its edge as a forecasting tool erodes every time a Polymarket probability proves more accurate than an analyst consensus. Proprietary research shops that sell “political risk” reports to hedge funds are watching their product get replicated, for free, by a website that lets you buy Trump shares for sixty-two cents.
This is the same pattern we’ve seen in every information industry that gets disrupted. Newspapers didn’t die because people stopped wanting news. They died because the distribution monopoly broke. Polling firms and data providers won’t die because people stopped wanting forecasts. They’ll struggle because the information asymmetry that justified their pricing is evaporating.
We’re not saying prediction markets are infallible. They’re not. The “French Whale”—a single trader who reportedly placed $85 million in bets on the 2024 election—showed that these markets can be moved by big money, not just big wisdom. Better Markets, a financial reform nonprofit, pointed out that Polymarket participants are not a statistically valid sample of voters, and that Kalshi and Polymarket sometimes show different probabilities for the same event—a situation that should be impossible in an efficient market.
There are also real regulatory questions. The CFTC amended its order on Polymarket in late 2025, and several states have opened court cases arguing it should be regulated as a gambling operator. The Atlantic Council has flagged prediction markets as a potential vector for foreign influence operations. These are not trivial concerns.
But here’s the thing: the same criticisms apply to the incumbents. Polls are manipulable (ask any campaign that’s ever commissioned an internal poll designed to generate a favorable headline). Analyst forecasts are biased (the sell-side has structural incentives to be bullish). And proprietary data is only as good as its methodology—which you can’t audit, because it’s proprietary. At least with Polymarket, the probabilities are transparent, the order book is public, and the incentives are aligned: you make money by being right, not by telling clients what they want to hear.
We did something very interesting… We had Claude Cowork, the newest Anthropic supertool, generate an entire financial model showing the TAM and potential market displacement caused by Polymarket. This took us thirty minutes. It would’ve taken an analyst, getting paid $100/hour or more, at least a day. Below are some screenshots we found interesting:




Jeez Claude, pretty impressive…
Prediction markets aren’t going to kill Bloomberg. They’re not going to make Gallup disappear overnight. But they’re doing something arguably more important: they’re democratizing access to the kind of forward-looking information that used to be locked behind paywalls and institutional gatekeepers. When the Wall Street Journal starts embedding Polymarket data next to its stock tickers, when DraftKings and Robinhood roll out prediction products for the 2026 midterms, when academic research consistently shows crowds-with-money outperform experts-with-surveys—the writing is on the wall.
The information advantage that defined institutional finance for half a century is getting compressed. The question isn’t whether prediction markets will reshape the data industry. It’s how fast the incumbents will adapt—or how slowly they’ll decline.
Thanks for reading & stay curious!
—J&E