A longtime operator of President Trump’s teleprompter has been placed on unpaid administrative leave after winning more than $100,000 on prediction market Kalshi based off the commander-in-chief’s remarks.
White House press secretary Karoline Leavitt told reporters Thursday that the operator, Gabriel Perez, would not be working the teleprompter for President Trump’s address to the nation on Thursday night.
“They’ll be on paid administrative leave,” she said before correcting: “I’m sorry, without pay. The administrative leave is unpaid. To be very clear, that was a decision by the president.”
Perez is speaking with the Commodity Futures Trading Commission to resolve allegations about bets he placed.
Follow The Post’s live coverage of President Trump and national politics for the latest news and analysis
He placed money on whether Trump would use specific words or phrases or address certain topics in his remarks. Perez did this for more than a dozen speeches, including February’s State of the Union address, January’s speech to the World Economic Forum in Davos, Switzerland, and a December primetime address, ABC News reported.
He is “fully cooperating with the CFTC on the matter,” a source familiar told The Post.
ABC News was the first to report the investigation.
Perez made more than $90,000 in profits from his trades, but the money has been frozen, CNN reported.
Kalshi is known as a “mention market,” where people can wager on words and expressions the president will or will not say during public events.
Prediction markets have found a huge audience in recent years, allowing people to be on anything from what color tie the president will wear to sporting outcomes.
They’ve also been ripe with instances of people using their insider knowledge to personally profit.
US Army Master Sgt. Gannon Ken Van Dyke was charged in April 2026 for allegedly using classified, nonpublic information about the top-secret military raid on Venezuela to win approximately $409,000 in prediction market bets.
He placed 13 bets totaling $33,000 on Polymarket between late December 2025 and early January 2026 about the operation that resulted in the capture of Venezuelan President Nicolás Maduro. Van Dyke has pled not guilty and is awaiting trial.
The case marked the first time the Justice Department filed insider trading charges involving a prediction market. The Commodity Futures Trading Commission also filed civil charges against Van Dyke.
Kalshi flagged the suspicious bets from Perez to the federal government.
“The Kalshi surveillance team promptly flagged, investigated and referred these trades to the CFTC,” Robert DeNault, head of enforcement and legal counsel at Kalshi, wrote on social media.
“We have been assisting regulators on this matter and provided all evidence that we collected, as we do with any referral,” he added.
A spokesperson for the CFTC declined to comment.
Perez is a longtime assistant to Trump and has operated the president’s teleprompter since 2016.
Given his position, he was one of the last people to see Trump’s speeches, getting a copy of the text to load in the prompter ahead of the president’s delivery. As a deputy assistant to the president, Perez was paid $175,000 a year, according to records.
Leavitt said Trump was unhappy with the matter.
“He believes it is deeply unfortunate and, frankly, a disgrace,” she said.
White House staffers, as government employees, face strict ethical guidelines when it comes to personally profiting from their positions.
“There are very strict ethical guidelines here at the White House that explicitly state not to do this, and the White House Counsel’s office makes that clear to all of us who sign up to work in government on behalf of the president,” Leavitt noted.
“This individual unfortunately violated the plan, and therefore he’s paying the consequences.”
She noted all staffers had to sign the ethics pledge as part of their onboarding process. Leavitt wasn’t sure whether Kalshi was available on White House systems or blocked.
The White House, in March, sent around a memo to staff warning them against using nonpublic information to place bets on prediction markets.
