What Are "Mention Markets"? Inside the Scandal That Triggered a CFTC Probe

In mid-August 2026, a strange story broke that put an obscure corner of the prediction-market world under a very bright light. President Trump's longtime teleprompter operator had reportedly been placing bets on Kalshi about which words the president would or would not say during public appearances — and using advance access to prepared remarks to do it. The account had built up more than $90,000 in profit before Kalshi froze it. Within days, the Commodity Futures Trading Commission (CFTC) opened a review into the entire category of contract behind the story: "mention markets."

What actually happened

According to reporting from NPR and several other outlets, the teleprompter operator had access to drafts of presidential speeches ahead of more than a dozen public events. Rather than treat that access as confidential, the operator allegedly used it to trade on markets asking whether Trump would say a specific word or phrase during the speech — essentially wagering on a script the operator had already read. Kalshi froze the account before the money could be withdrawn, but the episode raised an obvious question: how many other "mention market" positions rest on information the trader should never have had?

What is a "mention market"?

A mention market is a contract that settles on whether a specific word, phrase, or topic comes up during a defined event — a speech, a debate, a press conference, an earnings call, a broadcast. In principle it is not so different from any other prediction market: traders buy and sell shares in a yes/no outcome, and the price reflects the market's collective estimate of the probability. Kalshi has offered a range of these contracts on political and media events; Polymarket runs similar markets on its non-US site, though not on its CFTC-regulated US exchange.

The appeal is obvious. Mention markets are lightweight, fast-resolving, and tied to moments people are already watching live, which makes them fun to trade. The problem is what that same structure exposes: unlike an election result or an economic release, the "true" outcome of a mention market can be known in advance by a small number of people — the speechwriters, the staff, anyone who has seen the script. That is a very different risk profile from a market where nobody, including insiders, actually knows the future.

Why this is a manipulation problem, not just bad optics

Every prediction market has some version of this issue, which is why we have written before about what "insider trading" on a prediction market actually means. But mention markets sharpen the problem because the "inside information" is not a forecast or a judgment call — it is often a finished document sitting on someone's laptop. A market on "will a hurricane make landfall in Florida" cannot be manipulated by reading a script, because nobody has written the hurricane's script. A market on "will the president say 'tariffs' during this speech" absolutely can be, if the speech has already been drafted.

That distinction is exactly what CFTC officials pointed to when explaining the review: these contracts are "potentially very easy to manipulate," and the commission is taking a hard look at whether some of them should exist in their current form at all. It echoes the broader line the CFTC drew in its 2026 proposed rules for prediction-market contracts, which already separate markets that reward genuine forecasting from ones that mostly reward access.

The regulatory net is closing from more directions than one

The teleprompter episode landed in the middle of an already rough stretch for the big US platforms. On 13 August 2026, Baltimore's mayor and city council sued both Kalshi and Polymarket, arguing their sports-outcome contracts are unlicensed sports betting dressed up as financial products. New York's Attorney General filed a similar suit the previous month, and the New York City Council has opened its own investigation into marketing practices across several platforms. None of this is specific to mention markets, but it adds up to the same underlying theme: as prediction markets scale into the mainstream, the gap between "this looks like a forecasting tool" and "this looks like a bet with an information edge" is exactly where regulators are choosing to focus.

What it means if you trade

None of this means prediction markets are broken — it means market design matters. A market resolves fairly only when the underlying uncertainty is genuinely shared by everyone trading it. Election outcomes, sports results, and economic data releases are hard for any one participant to control. A scripted speech is not, and the mention-market episode is a useful reminder to ask, before you take a position: who could plausibly know this answer before I do?

It is also a good reason to build that instinct somewhere the stakes are not real money. On Outcomer, you can trade prediction markets with virtual funds, get comfortable reading prices and spotting where an edge is genuine versus where it might just be someone else's inside information, and build the habit of asking the right questions before you ever risk anything real.