Federal investigators confirmed this week that Gabriel Perez, the White House teleprompter operator responsible for displaying President Trump's prepared remarks, had placed more than $100,000 in wagers on prediction market platform Kalshi — betting, with documented success, on specific words and phrases the president would deliver in speeches Perez had himself loaded into the machine approximately forty-five minutes prior. The Commodity Futures Trading Commission is reviewing the trades. Officials described the case as "unprecedented," a characterization that required them to explain, at some length, what they believed a teleprompter operator's job was.
According to the CFTC referral reviewed by NPR, Perez had established a pattern of placing his bets in the window between receiving the final approved speech text and the moment the president approached the podium. This window, investigators noted, typically lasted between thirty and ninety minutes — ample time, in the judgment of federal regulators, to constitute an unfair informational advantage over the general public, who had not yet been handed the speech.
"The integrity of prediction markets depends on no single participant having material non-public information," said a spokesperson for the Bureau of Market Confidence, reached by telephone. When asked whether a man who physically controls what words appear in front of the president might structurally possess material non-public information as a condition of his employment, the spokesperson said they would need to refer that question to a different bureau.
Investigators further noted that Perez demonstrated what one filing described as "real-time adaptive trading behavior" — specifically, that he would adjust or hedge his positions mid-speech on occasions when the president departed from the prepared text. This, experts told govment.org, is less a scandal than an accurate job description. The president has departed from prepared remarks at a frequency that one White House logistical planning document, cited in the referral, classifies internally as "standard operating variance."
Kalshi, the prediction market platform at the center of the inquiry, received federal approval to operate election-related contracts in 2024 after a prolonged legal battle with the CFTC — the same agency now investigating trades made on its platform by a federal employee with advance knowledge of presidential statements. The CFTC declined to comment on whether this represented a conflict of interest, an irony, or both, noting that those determinations fall outside its current statutory mandate.
Legal scholars contacted for this article were divided on the core question of whether a man paid to know what the president will say should be prohibited from financially benefiting from knowing what the president will say. "You've essentially described the entire financial services sector," said one professor of securities law who asked not to be named, "but with a teleprompter." A second legal scholar said the case would hinge on the definition of "inside information" and whether it could apply to text the operator had not written, merely displayed, in a room he was permitted to enter because displaying it was his job.
Perez has not been charged with any crime. According to people familiar with the matter, investigators are still working through the foundational question of which law, precisely, covers a scenario in which the information asymmetry is the organizational chart. The White House did not respond to a request for comment. The teleprompter was unavailable.
The CFTC has announced it will convene a working group to study the matter and report its findings within eighteen months, at which point the administration whose speeches generated the trades in question will have approximately six months remaining in its term. The working group will be chaired by the Bureau of Looking Into Things.