The Company That Just Built the Price-Comparison Site for Prediction Markets Also Sells Data to One of the Venues It Rates

Genius Sports turned on Prediction.com on September 17. It does for prediction markets what Oddschecker does for sportsbooks: pulls the price of the same contract from eleven apps — Kalshi, Polymarket, Novig, Underdog, FanDuel Predicts and Fanatics among them — and stacks them side by side so you can see who is charging what.

In August, Genius Sports agreed to supply Kalshi with official data for major soccer competitions including the Premier League and Serie A.

Those two sentences describe the same company in the same quarter. The first is a referee. The second is a vendor to one of the players. Nobody hid either one — the data deal was announced, the launch was announced — but they have not been presented together, and the comparison page does not tell a visitor that the company running it has a commercial relationship with one of the venues on the grid.

What was actually bought, and for how much

Prediction.com is built on Legend, and the price of Legend is the part of this story that explains the rest of it.

Genius Sports' own filing with the SEC, dated February 5, lays out the terms: $800 million in cash and $100 million in stock at closing, plus an earnout of up to $300 million paid over the two years after close and tied to profitability and cash-flow thresholds. Up to $1.2 billion in total. Genius funded it with an $850 million Term Loan B and said pro forma leverage would sit below 3.0x. The deal was expected to close in the second quarter.

What that money bought was attention: 320 million visits from 118 million unique visitors in 2025, with more than two-thirds of those being returning visitors. Mark Locke, Genius Sports' chief executive, described the transaction as one that "supercharges fan monetization, and builds a fully integrated sports and gaming media network." Legend founder Nick Kisberg talked about unlocking growth "for our partners and products."

Neither man said "affiliate," but that is the business. A comparison site makes money by sending a user to a venue, and the venue pays for the user. The interesting question about any such site is never whether it is accurate. It is what happens when accuracy and the payment point in different directions.

The tell is the combination engine

The feature on Prediction.com worth staring at is not the side-by-side single-contract grid. It is the multi-leg combination analysis — pricing the same parlay-equivalent across different venues.

That is where the margin lives, and this desk has spent the month showing it. Single-game prices on prediction markets have genuinely compressed; the reason a combination engine is the headline feature is that combination pricing has not compressed, and the dispersion between venues on an eleven-leg ticket is enormous compared with the dispersion on a moneyline. A tool that surfaces that dispersion is genuinely useful to a trader. It is also, for precisely the same reason, the most commercially valuable page on the site, because it is the page where the user is about to do the highest-margin thing they will do all week.

The pricing tiers confirm what kind of business this is. There is a free consumer tier, a developer package at a million monthly comparison requests, and a professional tier at up to five million. That is not a media property with a rate card. That is a data vendor with a consumer front end.

Why the data deal matters more than it looks

Take the charitable reading first, because it is the correct starting point. Genius Sports sells official league data to essentially everyone. It has supply relationships across the entire betting industry. If a data contract with one venue disqualified the company from ever building a comparison product, Genius could never build anything, and the same logic would bar Sportradar and every other supplier. Data supply is plumbing. Plumbing is not endorsement.

That reading is strong and it is probably how Genius Sports would answer this piece. Here is where it runs out.

The comparison layer is not a neutral utility, because the thing being compared is price, and Genius Sports has a commercial interest in the volume of one of the venues whose price it displays. Kalshi pays Genius for data. Kalshi's payment capacity scales with Kalshi's volume. Prediction.com influences where volume goes. The conflict does not require anyone to tilt a number; it only requires ordinary product decisions — which venue is listed first, which is in the default comparison set, which gets the combination engine's deepest coverage, which one the editorial content on a Legend property happens to explain most clearly — to break, on the margin, in a direction that is also good for the data contract.

This is not a hypothetical failure mode. It is the exact failure mode that hollowed out sportsbook affiliate comparison in the first place, where "best odds" pages reliably ranked by commission rather than by price, and it took regulators and a decade of reader distrust to make that visible.

The disclosure this story requires of us

Two of them, in fact.

TrueEdge builds odds and pricing tools and earns affiliate commissions from sportsbook partners. Prediction.com is, in the most direct sense available, a competitor — a far better capitalised one — and we would have an incentive to be unkind to it even if it were flawless. Readers should discount accordingly.

The second is more specific and more uncomfortable. Legend, the network Genius Sports bought, owns Covers.com and Casino.org. We have cited Casino.org in this publication within the last week, including in our reporting on handle forecasts. That does not make those citations wrong, and we are not retracting them. It does mean that a company we are writing critically about today owns outlets that appear in our sources blocks, and you should know that before you weigh either.

What we do not know

We have not been able to confirm the financial terms of the Genius–Kalshi data agreement, or whether it is fixed-fee or volume-linked. That distinction matters enormously to everything above: a flat annual licence creates a weak conflict, while a per-event or revenue-share structure creates a direct one. We asked the question of the public record and the public record does not answer it. Until it does, the honest statement is that the conflict is structural and unquantified, not that it has been exercised.

We also do not know whether the eleven-app list is editorially selected or simply everything with a usable public feed. Those produce identical-looking pages and very different products.

The larger point

Comparison layers do not appear early in a market. They appear when a product has become a commodity and the only remaining question is price — which is why Oddschecker exists for sportsbooks and why nothing like it exists for, say, streaming rights. Genius Sports spending up to $1.2 billion and then pointing the asset at prediction markets within months is the clearest institutional signal yet that the smart money thinks event contracts are now a price business rather than a product business.

The exchanges have spent two years arguing they are something categorically different from a sportsbook. The first serious infrastructure built on top of them is a shopping engine.