There is something deeply funny about using the Oscars to advertise the power of prediction markets.
To be clear, the Academy does not deal in corn futures. It is a vote conducted by 10,000 members of the global entertainment elite, scattered across the world, and influenced by taste, status, relationships, fashion, momentum, and mood. And yet the fintech guys would like us to believe this was a triumph of price discovery.
About $120 million was wagered on the 2026 Oscars as prediction markets went mainstream. Supporters of platforms like Polymarket and Kalshi argue that when markets reach enough activity, their collective wisdom becomes a kind of crystal ball — faster than experts, smarter than pundits, and better at identifying outcomes before the rest of the culture catches up. Their evidence: prediction markets correctly called 19 of the 24 Oscar categories.
That is the sales pitch. The reality is less impressive.
The sales pitch
Prediction markets insist they are not gambling. They describe themselves in the language of finance, not sportsbooks. Users buy and sell “event contracts” with one another rather than betting against the house — a casino, FanDuel, or a bookie — that takes the other side of the wager.
That distinction matters because, positioned as financial instruments, these companies can sidestep some of the scrutiny applied to gambling businesses. Kalshi board member Alfred Lin, a Sequoia Capital partner whose firm was an early investor, has argued that these markets serve a legitimate economic purpose by allowing investors to hedge against uncertainty — fluctuating interest rates, for example, or a geopolitical crisis that disrupts oil supply.
How that logic extends to multimillion-dollar Oscar markets on which celebrities would appear at the Oscars is less clear. It also may not matter. In January, guidance from the U.S. Commodity Futures Trading Commission helped open the floodgates, and Kalshi’s valuation reportedly jumped from $1 billion to $11 billion in three months.
The idea that these companies are meaningfully different from gambling sites — or, more to the point, that their growth is not being driven by young men gambling on their phones — is hard to take seriously. Fortune reported that 90 percent of recent volume in these markets appears to stem from traditional gambling behavior.
What began as an “experiment” in sports and entertainment has quickly become central to the business. This year’s Super Bowl alone generated $1 billion in event contracts, and Kalshi’s reported revenue jumped from $1.8 million in 2023 to $24 million in 2024, with estimates of $260 million in 2025.
The intellectual case for prediction markets was always more ambitious. Academics, investors, and data journalists like Nate Silver argued that highly liquid markets — where contracts can be bought and sold continuously, like securities — might produce powerful new predictive signals. In theory, they should react faster than pundits or pollsters and emerge as early indicators of cultural, political, and financial shifts.
Lin called prediction markets “truth machines.” Kalshi cofounder Tarek Mansour put it this way: “They’re a market-based mechanism, so you get the wisdom of the crowds… when people have real money on the line, they don’t lie.”
President Trump’s 2024 election victory was widely framed as the breakthrough moment for prediction markets, with roughly $3 billion placed on the outcome. But that election was also a prime example of how overstated their predictive power can be.
Between October 29 and November 3, 2024, Trump’s odds of winning on Kalshi fell from 64.2 percent to 50.8 percent. Then, between 8 p.m. on November 4 and Election Day on November 5, they surged to 66.7 percent. That is not a market acting as an early-warning system. It is a gambling market absorbing the latest polls, punditry, and momentum narratives as last-minute money pours in.
Following the pundits
The same dynamic played out with the 2026 Oscars. You may have seen headlines claiming prediction markets had a banner night, with Kalshi performing about as well as top Oscar pundits like IndieWire’s Anne Thompson, the Next Best Picture team, and The Hollywood Reporter’s Scott Feinberg. It correctly picked 18 of the 21 non-shorts categories.
That framing is misleading. The market’s confidence only emerged in the final 72 hours, after dramatic swings that closely tracked the experts’ final calls. The notion that the market somehow got an early read on the tastes of Academy voters is absurd. This is one of the most socially coded and insular electorates in public life. There is no clean predictive model for it because there is no real equivalent.
What these late-moving markets were actually following was the collective judgment of the reporters and analysts covering the Oscar race. That reporting helps decode one of the most closed systems in Hollywood, which makes it all the more ridiculous for a fintech company to use the Oscars as proof of its disruptive power.
An unhealthy market
I monitored Kalshi during the final six weeks of the Oscar race, and what I found was not a healthy predictive market. It was an unhealthy gambling market.
To see the difference, consider a normal sportsbook. Say Vegas sets the point spread for the New York Giants at -3.5, meaning the Giants must win by four points for that bet to cash. If the Giants are coming off a dominant performance, retail gamblers may overreact and push the line to -6.5.
That movement creates opportunity. A sportsbook wants balanced action on both sides, so distorted odds attract bettors in the other direction. Sharp money steps in when the public has overreacted. Professional gamblers look for those moments because, over time, consistently exploiting pricing errors is how they make money.
And if the line later corrects, those same bettors may place a hedge in the opposite direction. A healthy gambling market keeps self-correcting this way, balancing herd behavior against professional skepticism.
Kalshi during Oscar season behaved nothing like that.
Historically, Oscar favorites win about 70 percent of the time, though that fluctuates from year to year across a very small sample of 24 categories. And not all favorites are alike: in competitive races, the leader often remains unclear until precursor awards like SAG, PGA, and DGA.
This year’s Best Actress and Best Actor races show how badly Kalshi handled both a relatively settled contest and a volatile one.
In Best Actress, eventual winner Jessie Buckley was about as close to a lock as the race offered. But on Kalshi she rose to an absurd 98.6 percent favorite. The problem is not simply that the odds looked silly. It is that no Oscar race contains enough information to justify that level of certainty.
Precursor awards are less predictive than they appear. Buckley won SAG, yes, but we do not know by how many votes. Without underlying vote totals, it is difficult to build a robust model or even assess how much signal the result contains. And the overlap between SAG’s 160,000-member union and the Academy’s much smaller acting branch is tiny.
So yes, Buckley was the favorite. But 98.6 percent implies a level of certainty the underlying data simply cannot support. And Buckley was not an isolated case. Olympic figure skater Ilia Malinin reached a 99.1 percent chance of winning on Kalshi — a bizarre number for a sport where one athlete, under immense pressure, performs a three-minute routine on ice against the best skaters in the world. Who is taking that risk to make a penny on the dollar?
That question matters because it suggests these markets are not being disciplined by rational price discovery. They are being distorted by users treating them as gamified betting apps, often in combination with parlays and other high-risk wagers.
The Best Actor race made the distortion even clearer.
Before the SAG Awards on March 1, Timothée Chalamet was about a 72 percent favorite on Kalsia. In practical terms, that meant you could buy a 28-cent contract on any one of the other four nominees winning. Didn’t matter if it was Michael B. Jordan, Ethan Hawke, Wagner Moura, or Leonardo DiCaprio; anyone but Chalamet wins, and you collect a dollar on your 28-cent bet.
But the data never justified Chalamet as such a strong favorite in a race pundits repeatedly described as unusually open. Jordan and DiCaprio were beloved stars from the two biggest films positioned to dominate the Oscars. Ethan Hawke’s one-man showcase in “Blue Moon,” combined with a strong campaign trail charm offensive, made him a legitimate contender. And Wagner Moura, already a Golden Globe and Cannes winner, was earning raves for “The Secret Agent,” making him a viable favorite amongst the Academy’s increasingly international membership.
Then, less than two weeks later, after Jordan won the SAG awaard, Chalamet contracts were trading at 33 cents on the dollar.
Put differently, within less than two weeks, Kalshi offered both of these bets:
$100 on Chalamet losing: payout $357.14
$100 on Chalamet winning: payout $303.03
Yes, close races move. But this kind of swing is nearly impossible to replicate in a healthy, high-volume sportsbook market. What we are seeing here is not the wisdom of crowds. It is the gamification of every news event, turning terminally online behavior into the illusion of edge and channeling that impulse into a flood of money.
Worse, prediction markets are increasingly adopting the same predatory mechanics as casinos and sportsbooks.
In early February, Bank of America estimated that nearly one-fifth of activity on Kalshi and Polymarket came from parlays, wagers that bundle multiple bets together. For example: A single bet that Buckley, Chalamet, and “Golden” all win Oscars. If all three hit, the payout is larger. But the larger payout never fully compensates for the added risk; the math always favors the house. Anyone who has struggled with gambling, or known someone who has, understands how predatory parlays can be.
Prediction markets justify this by wrapping parlays in the language of derivatives and trading, as Fortune noted. That may also explain the irrational appetite for extremely short-money bets like Buckley at 98.6 percent. On their own, those bets make little sense. Bundled into multileg combinations, they begin to look much more like casino products dressed up as financial instruments.
Call it gambling, call it prediction markets, call it derivatives trading — increasingly, it looks like Wall Street’s newest way of monetizing what it lovingly calls dumb money.
