Kalshi's Data Was Public Until Someone Read It

The window was open for less than 48 hours.

Dune, a company that provides tools to track prediction market activity, pulled two Kalshi datasets from public view on May 15. The timing was not subtle: the removal came less than 48 hours after Sportico used that same data to analyze how much retail bettors were losing on parlays. The datasets didn't quietly expire or get quietly reorganized. They went behind a $40,000 paywall. Dune didn't acknowledge the change until the matter was flagged.

A writer at Sportico.com staked out the obvious implication. And the obvious implication deserves to sit with you for a moment.

What Transparency Looks Like Before It Gets Expensive

Prediction markets have spent considerable energy presenting themselves as something different from traditional sportsbooks — more transparent, more data-forward, more honest about the mechanics. That argument depends on the data actually being visible. Visible to researchers, to journalists, to the bettors themselves. The pitch only works if the ledger is open.

What Sportico's piece reveals — not by arguing it, but simply by existing — is that the ledger was open until someone read it. And what they read, apparently, was not flattering to the house.

We don't know exactly what the parlay loss figures showed. The source doesn't detail the numbers beyond establishing that the analysis covered retail bettor losses. But we know what typically lives inside parlay loss data, and we know what it looks like when a company's first instinct, upon that data being published, is to lock it away behind a price that filters out everyone except institutional players.

$40,000 is not a fee for access. It's a fee for silence.

The Architecture of Selective Openness

This is worth naming clearly: Dune didn't create the data. Kalshi is the prediction market platform. Dune provided the tracking infrastructure. The datasets were public — that was a choice, presumably a marketing choice, the kind of openness that looks good in a pitch deck and reassures regulators who are still figuring out what prediction markets even are.

Then a journalist used the infrastructure the way infrastructure is supposed to be used. She found something. The data moved behind a wall.

The sequence matters more than any single number in the original report. Because the sequence tells you what the actual policy is, as opposed to the stated one. Transparency that disappears when it becomes inconvenient isn't transparency. It's a prop.

Prediction markets are relatively new terrain for regulators, for bettors, and for the sports media ecosystem trying to cover them. Kalshi and companies like it have benefited from the uncertainty — from the sense that this is a different category of thing, that the old critiques of gambling don't quite apply here. That benefit of the doubt was never unlimited. It was borrowed against the promise of openness.

Spending that promise to protect a 48-hour news cycle is a peculiar trade.

The $40,000 number will be forgotten. The pattern won't be.