Kalshi put money on Byron Donalds twice. First as a donor. Then as a bookie. The prediction market that wants to be treated as a financial exchange, not a sportsbook, gave $25,000 last year to the political committee for the Florida congressman now running for governor. Then it went on social media and advertised his odds at a “NEW ALL-TIME HIGH.” Donalds reposted it and wrote “wow.” His campaign called the post “reality.” Kalshi’s spokeswoman said the company posts about every candidate when the numbers move.
What they did not say is the obvious part. The house donated to the horse, then sold the public a price that treats him losing as a rounding error.
The Price Is the Tell
Kalshi’s own site has been pricing Donalds around 98 percent to win the Republican nomination. Fishback is a penny. Collins is less than a penny. Polymarket has been even more aggressive, putting Donalds at 99 percent. That is the number they want voters, donors, and journalists to treat as a forecast. It is not a forecast. It is a sales pitch with a ticker.
Donalds is the favorite. That is not in dispute. He has Trump’s endorsement, the Cabinet, Rick Scott, and a war chest his rivals cannot touch. Ken Griffin dropped $10 million on him. The Tampa Bay Times average has had him up by roughly 40 points. A candidate in that position should be favored. He should not be priced as if the remaining risk is one percent.
The polls do not say that. Quantus had him at 50 percent against Collins at 16 and Fishback at 12. Targoz had him at 49 to 18. The Political Matrix poll had him at 32, Renner at 19, Fishback at 13, and 32 percent still undecided. Even the campaign’s own Fabrizio numbers put him in the low 50s. A man sitting at 50 percent in a multi-candidate primary, with a sitting governor calling him unfit and refusing to hand him the job, is not a 98-cent contract. Ron DeSantis has spent months arguing that Donalds should not replace him. He has said it publicly. He has said it to Trump. That is not a one-percent problem.
The general election is worse. Change Research had David Jolly up on Donalds 46 to 42 in May. UNF later had Donalds up 46 to 41. Targoz had him up 45 to 38, and also had Collins beating Jolly by more. Independents in that survey broke against Donalds by 17. Kalshi’s general-election market still priced the Republican nominee around 81 percent. Florida is red. It is not a coin flip. It is also not an 81-percent lock for a nominee whose own party’s outgoing governor will not stand next to him.
If Donalds’ true chance of losing the primary is even 5 or 10 percent, the market is lying. Buying him at 98 cents means risking 98 dollars to win 2. One loss wipes out 49 wins. That is not a market discovering truth. That is a market selling certainty to people who do not know how to read a price.
They Bought the Regulators Too
The $25,000 to Friends of Byron Donalds was not a one-off act of civic virtue. Kalshi also sent $50,000 to Attorney General James Uthmeier’s committee and another $3,000 to his campaign. It gave $25,000 apiece to the Republican House and Senate campaign committees in Florida, $25,000 to a committee run by House Speaker Danny Perez, and $5,000 to a committee run by the next speaker, Mike Redondo. The same company that wants Florida to treat it as a federal exchange, not a casino, has been writing checks to the people who would decide that question.
That is the actual business. Florida’s gambling compact gives the Seminole Tribe the sports-betting franchise. If Kalshi is gambling, it is trespassing. If Kalshi is a CFTC exchange, it gets to take bets on elections and football without the tribe, without the state book, and without the rules that apply to DraftKings. The Trump administration has been suing states that try to regulate these platforms as gambling. Kalshi needs friends in Tallahassee while that fight is live. So it bought some.
Michael Beckel at Issue One said the quiet part. People can be betting a race that the exchange itself has already wagered on with campaign cash. Most users never see that. Kalshi posted Donalds at 95 percent in June without disclosing the donation. The company later said it posts about Fishback and Collins too. Posting about the field is not the same thing as telling customers the house funded the favorite.
Kalshi has run the same play in California, putting about $40,000 behind Democrat Xavier Becerra, who also leads its market there. This is not ideology. It is customer acquisition and regulatory insurance. The candidate gets money. The exchange gets a friend. The market gets a price that looks like destiny. Everyone else gets the sucker side of the trade.
Worse Than a Sportsbook
Sportsbooks may be parasitic. They are also honest about the job. And sports betting is also the least profitable form of gambling for the house. And they stand to lose money on many props where the sheep correctly predict an outcome. Thus they almost always advertise parlays and creative betting rather than point spreads and over-unders.
A sportsbook posts a line, takes both sides, and lives on the vig. The standard juice on a spread is about 4.5 percent. You know the tax. You know the rules. A touchdown is a touchdown. The ball either crosses the plane or it does not. DraftKings does not write a $25,000 check to the Chiefs, tweet that Kansas City is a 98-percent lock, and then reserve the right to settle the game off a clause on page 140. The house is not trying to swing the result. The result is on the field. The book wants action on both sides so it does not care who covers.
Prediction markets sell the opposite story. They tell you they are a truth machine. Wisdom of crowds. Price discovery. A better poll. Then they charge you to trade, let thin order books get shoved around by a few large accounts, and keep a rulebook long enough that the house can decide what “happened” after the fact.
Kalshi’s trading fee runs about 1 percent per side, with a floor and a cap. On a 98-cent contract, the fee is a joke at your expense. The posted edge is already gone before the election is called. Sports betting at least offers plus-money on dogs that actually hit. Buying a politician at 98 cents is paying favorite’s tax for the privilege of collecting pennies, while the only contract with a real payout is the one the exchange has already crushed.
The ROI is worse because the product is worse. A sportsbook’s worst day is a bad beat. A prediction market’s worst day is the exchange announcing that the thing you watched happen does not count.
The House Writes the Fine Print
Kalshi users agree to more than 180 pages of terms, rules, and policies. That is over 67,000 words, plus a separate rule sheet on every contract. Diehard traders treat the footnotes as the real market. Everyone else clicks through and assumes the title of the bet is the bet. It is not.
When Israel and the United States killed Iran’s Supreme Leader, Kalshi had a market on whether Ali Khamenei would be “out.” Users thought dead counted. Kalshi settled off the odds from one minute before the bombs fell, then spent $2.2 million cleaning up the complaints. One trader called it a bait-and-switch and emptied his account. Kalshi said people might not have seen the rules because of how the app is designed. The company later filed a new death rule: if the person at the center of a market dies, related bets pay out at the last price before death was “reasonably anticipated.” The title can say one thing. The settlement sheet says another.
That is the product. The house does not have to buy a candidate to tilt a market, though Kalshi did that too. It can let a thin book get walked up to 98 cents. It can promote that number as news. It can rewrite the resolution criteria when reality is messy. The platforms answer with more surveillance language and more pages. The rulebook gets longer. The customer still thinks he bought “Donalds wins.”
A sportsbook cannot do this. The NFL does not let FanDuel decide after kickoff that a field goal was actually a void. The league can be corrupt. Refs can be blind. The book is still betting the scoreboard, not a memo it drafted after the whistle. Prediction markets collapsed that distinction on purpose. They want the cultural status of a forecast and the legal status of a derivative and the payout discretion of a carnival booth.
What the Number Is For
A 98-percent Donalds price does work. It just does not do the work they advertise. It tells low-information bettors the race is over. It tells reporters they can write the nomination as a formality. It tells donors they would be fools to fund anyone else. It tells Tallahassee that the same company writing checks to Donalds, Uthmeier, and the legislative leadership is also the company publishing the “market” that says resistance is irrational. The price becomes a campaign ad that never has to have an authorization line.
Donalds may win the primary. He is the favorite, and favorites win most of the time. That is not the point. The point is that a platform with a cash position in his operation is selling the public a one-percent chance he loses, collecting fees on the way up, and keeping a rulebook that lets the house settle disputes in its own favor. Sports betting is a tax on people who think they can beat a spread. Prediction markets are a tax on people who think they are participating in a forecast, when they are sitting at a table where the dealer donated to one of the hands and wrote the definition of a win.
If a sportsbook posted the Chiefs at 98 cents in Week 3, nobody would call it wisdom. They would call it a sucker price. Kalshi posted that price on a politician it funded, then asked to be regulated like the New York Stock Exchange. The receipts are clear. The house is not discovering the future. The house is selling it.
