A veteran White House insider has been ousted after using the president’s own speeches to line his pockets through online prediction markets. Gabriel Perez, Donald Trump’s longtime teleprompter operator, is under federal investigation for allegedly betting tens of thousands of dollars on the exact words the president would say during live broadcasts. The scandal has exposed a massive vulnerability in the rapidly growing world of political wagering, proving that the greatest threat to market integrity is not foreign interference or sophisticated algorithms, but the low-tech advantage of a staffer with a USB drive and a front-row seat.
This is the first time an active White House staff member has been caught exploiting privileged access for prediction market profits. It marks a significant escalation in a trend that financial regulators have struggled to contain. On Thursday, White House Press Secretary Karoline Leavitt confirmed that Perez had been placed on unpaid administrative leave and is no longer working at the executive mansion, calling the situation deeply unfortunate and a disgrace.
Behind the scenes, the Commodity Futures Trading Commission is negotiating a settlement with Perez, who has reportedly been cooperating with investigators after the trading platform Kalshi flagged his trades and froze nearly $100,000 in illicit winnings.
The Anatomy of a Speech Betting Scheme
For nearly a decade, Gabriel Perez was one of the most trusted behind-the-scenes figures in the Trump orbit. Hired almost by accident during the 2016 campaign when Trump’s team searched Google for local teleprompter services, Perez rose to become the Deputy Assistant to the President and Technical Adviser. He was the sole individual Trump trusted to sync the scrolling glass screens with his unpredictable, highly ad-libbed speaking style.
This unique level of trust granted Perez early access to speech drafts. While high-profile addresses like the State of the Union or speeches at the World Economic Forum are guarded by tight internal security, the teleprompter operator must receive the final script to format it for the scrolling hardware.
Perez realized this logistical bottleneck was a goldmine.
On prediction platforms like Kalshi, traders buy and sell contracts on "mention markets". These micro-markets allow users to bet on whether a political figure will utter specific words during a public appearance. If a contract asks whether Trump will say the word "tariffs" during a prime-time address, a trader can buy "Yes" or "No" shares.
According to investigators, Perez began loading up on contracts that perfectly matched the drafts on his screen. He allegedly placed bets on commonplace words, economic terms, and specific country names across more than a dozen speeches.
What made his strategy even more brazen was his live trading. Because Trump is notorious for veering off-script, Perez reportedly monitored the live speeches with his betting app open. If Trump skipped a pre-written paragraph containing a word Perez had bet on, he would quickly exit his position or hedge his bets while the president was still speaking on national television.
The scheme worked perfectly until the platform's internal security noticed a pattern that was simply too accurate to be luck.
The Rise and Threat of Mention Markets
The concept of a prediction market relies on the wisdom of the crowd. When thousands of independent actors trade on the likelihood of an event, the price of the contract becomes a highly accurate probability. However, when a market is based on something as easily manipulated as the spoken word, the crowd’s wisdom is easily defeated by a single person with early access to the script.
Mention markets are a relatively new and highly controversial product. While platforms like Polymarket and Kalshi built their reputations on macroeconomic indicators and election outcomes, they introduced micro-wagers to drive engagement. Users can bet on whether a CEO will mention a competitor, or if a press secretary will use a specific phrase.
The vulnerability of these markets was publicly mocked last year by Coinbase CEO Brian Armstrong. During an earnings call, Armstrong acknowledged he was watching a Polymarket wager regarding his speech. He jokingly rattled off a list of words, including bitcoin, blockchain, and Web3, specifically to trigger payouts for the traders.
While Armstrong’s stunt was a lighthearted critique of the system, the Perez case represents a much darker reality. It shows that anyone with access to sensitive communication can treat prediction markets as a personal ATM.
Major brokerages have already warned about these structural flaws. Robinhood, which recently partnered with Kalshi to offer event contracts, intentionally excluded mention markets from its platform due to concerns over manipulation and insider trading. The Perez scandal validates those concerns, demonstrating that even the highest offices in the land are not immune to the temptation of a guaranteed payout.
The Legal Grey Area of Prediction Market Insider Trading
The CFTC is now faced with a complex legal challenge. While insider trading laws are highly developed for traditional stock and commodities markets, their application to prediction markets is still being tested in the courts.
Traditional insider trading requires a breach of fiduciary duty involving material, non-public information that impacts the value of a security or commodity. Do speech drafts count as commodities?
Under the Commodity Exchange Act, the CFTC has broad authority to police fraud and manipulation in contracts traded on registered exchanges like Kalshi. The agency is expected to argue that Perez misappropriated confidential government property—the president’s speech drafts—for personal financial gain, which constitutes a form of wire and commodities fraud.
The federal government has already signaled its willingness to prosecute these cases aggressively. Earlier this year, federal prosecutors indicted Gannon Ken Van Dyke, a U.S. Army soldier who allegedly made over $400,000 on Polymarket. Van Dyke allegedly used classified intelligence regarding a planned military operation in Venezuela to bet on the outcome of the mission.
The Van Dyke prosecution and the Perez investigation show that the Department of Justice and the CFTC are no longer treating prediction markets as harmless internet novelty sites. They are treating them as real financial markets subject to federal anti-fraud laws.
For Kalshi, the scandal is a double-edged sword. On one hand, its internal compliance team successfully detected the suspicious activity, froze the funds, and referred the case to regulators. On the other hand, the incident highlights how easily their markets can be compromised by a single bad actor.
The Structural Fix Campaigns Must Adopt
As prediction markets grow more liquid, the incentives for insider trading will only increase. If a single teleprompter operator can net six figures on a few minor speeches, the potential payouts for leaking sensitive macroeconomic policy changes, corporate earnings, or regulatory decisions are astronomical.
To survive, prediction markets must implement stricter controls. Kalshi recently announced it would require employment disclosures for high-risk markets and launch a formal whistleblower portal. However, self-regulation may not be enough to satisfy federal watchdogs.
On the administrative side, the White House and political campaigns will have to change how they handle draft remarks. Historically, speeches are treated as political secrets, leaked early only to allied journalists for favorable coverage. Now, they must be treated as highly sensitive financial information.
Campaigns may need to restrict draft access to an incredibly small circle, format the teleprompter files at the absolute last second, or even introduce deliberate, randomized changes to speeches to foil anyone attempting to trade on the text.
The era of treating presidential speeches as simple political theater is over. In a world where every word has a price tag on a betting board, a speech is no longer just a message to the public. It is a financial instrument.
