MLB sold prediction-market exclusivity for $75m a year to a defendant in roughly 20 states

Divide the MLB headline by its term and you get $75 million a year ^[10], or about $205,000 for every day the four years run ^[11]. That is the rate at which contracted value accrues on top of an unanswered question about which regulator owns the product: an event contract under the CFTC, or a sports wager under a state gambling commission, with the platforms arguing the first and a widening group of state attorneys general arguing the second ^[8]. Twenty or so states have put that argument into a filing already ^[2].

What the reporting does not give us is the term sheet, which is where the answer actually lives ^[14]. Whether the $300 million is guaranteed, whether it steps down as a platform geofences hostile states, whether an injunction counts as a termination event or merely a bad quarter for a partner: each of those turns the same nine-figure headline into a different instrument. A four-year guarantee with no regulatory out is credit exposure to a litigant. A four-year number with an out is a one-year deal wearing three renewal options.

The structural point is the exclusivity. MLB sold one venue the whole league ^[1], MLS did the same thing in January 2026 ^[4], and Kalshi went and bought five individual clubs instead ^[5]. Seven MLB teams now carry some prediction-market affiliation, which means two of them arrived by a route other than Kalshi's five branding deals ^[12]. If Polymarket loses in a meaningful set of states, the league-level seller has nowhere to move the category, because it sold the category; the team-by-team buyer holds five small positions that fail independently. Cryptobriefing makes a softer version of this point, that team-level deals could prove more resilient if league-wide ones are challenged ^[15].

This is probably wrong, but the data leg looks like the part that survives. Sportradar's expanded Polymarket arrangement covers more than 20 leagues and roughly 300,000 matches a year ^[6], and leagues are paid for supplying official data inside these structures ^[9]; a data licence does not obviously die because a venue has to register differently in one state. Or rather, the more interesting version: if the platforms end up licensed state by state as sportsbooks, the leagues have re-signed sportsbook partners at prediction-market prices, and the data pipe is the only piece whose price was never a function of the definitional argument.

This could resolve three different ways. Federal jurisdiction holds, and $75 million a year for exclusivity in a category that went from the NHL's October 2025 deal to the August 27, 2026 Sportradar announcement in about ten months ^[13] looks cheap. The states win partially, the platforms geofence, and whatever revenue share sits behind the guarantee never triggers. Or it settles into licensing, and the premium for being a market rather than a book compresses toward zero. The memorandum of understanding with the CFTC ^[7] reads as the leagues buying an option on the first outcome, and cryptobriefing takes the agency's willingness to sign it as accommodation after years of caution on sports event contracts ^[16].

What would prove this wrong: a ruling in any of those suits that the CFTC's jurisdiction over these contracts is exclusive, which would take the counterparty risk to near zero and make $75 million a year look like an underpayment. Equally, disclosure that the $300 million is terminable annually at regulatory discretion would cut four years of exposure down to one, and the number that would matter would be the annual $75 million, not the four-year $300 million.