Original | Odaily Planet Daily (@OdailyChina)
Author | Golem (@web3_golem)
Recently, another insider trading scandal has been exposed at the White House.
A White House staff member profited hundreds of thousands of dollars by trading on insider information in a prediction market. The identity of this insider was just a long-term teleprompter operator for Trump's speeches. The employee has now been suspended without pay.
This teleprompter operator has become the third insider disclosed by the US judicial department, following a special forces soldier involved in the Maduro capture operation and a Google security engineer, who made large profits in prediction markets using insider information.(Related reads: 《After 4 Months, Polymarket Helped Trump Catch the Leaker of the Military Operation, But at the Cost of...》《Looking at the Answers Before Handing in the Paper? Google Engineer Caught Up in Polymarket Insider Trading Case》)
Reported by Kalshi, Funds Frozen, but Ultimately No Criminal Liability
The protagonist is named Gabriel Perez, who has been responsible for operating the teleprompter for Trump's speeches since 2016. Perez's journey to this job is quite dramatic. In 2016, Trump's campaign team urgently needed a teleprompter operator. When they searched for "teleprompter" on Google, they found Perez's company, and Perez was hired by the Trump team just like that.
Gabriel Perez
Although Perez was hired by chance, over these 10 years, he gradually became one of Trump's closest aides. American "Politico" even stated that "Perez has become the only person Trump trusts." He often receives last-minute revisions to public speeches from Trump himself.
Therefore, Perez became one of the few people who could obtain Trump's complete speech drafts in advance and virtually had the final say on almost all of Trump's prepared speech drafts. This power is not insignificant. Perez's official title at the White House is Deputy Assistant to the President and Technology Advisor, with an annual salary of $175,000, only $20,000 less than senior staff like Chief of Staff Susie Wiles and Press Secretary Caroline Levitt.
Such a salary is already considered high-income in the US, but the greedy Perez was still not satisfied.
When prediction markets became popular, countless players began betting on which specific words Trump would "mention" in a certain speech. Perez discovered that his "privilege" could bring him even more wealth.
CFTC investigators found that over about three months, Perez placed bets on over a dozen of Trump's speeches, making a total profit of over $100,000. This included Trump's primetime speech last December, his speech at the World Economic Forum in Davos, Switzerland in January this year, the State of the Union address in February, and Trump's speech at the Medal of Honor ceremony in March.
The US President's statutory annual salary is $400,000. With various allowances, the President receives about $569,000 annually. If Perez hadn't been caught, at his rate of earning $100,000 in 3 months, although his power is less than the President's, his annual income would exceed the President's salary.
However, even knowing the speech content in advance, Perez couldn't always successfully predict which words Trump would mention in his speeches, because Trump often deviates from the script for "off-the-cuff" remarks. When Trump skipped a word Perez had bet on during a speech, Perez would immediately sell to cut his losses. Trump himself admitted in a speech at the Detroit Economic Club in January that 80% of the time he doesn't look at the teleprompter.
Just like the special forces soldier and the Google security engineer, Perez's exposure also stemmed from the prediction market platform's active reporting. Perez frequently used Kalshi for insider trading. Starting in March this year, Kalshi's monitoring system detected some abnormal trades related to specific words mentioned in Trump's speeches and thus noticed Perez.
After concluding its internal investigation, Kalshi quickly froze over $90,000 in Perez's account and handed the case over to the US Commodity Futures Trading Commission (CFTC). Upon learning of this, Trump commented that it was "despicable" and personally decided to suspend Perez without pay during the suspension.
Ultimately, Perez's greed cost him dearly. Not only did he fail to keep his prediction market profits, but he also lost his original job. However, compared to the special forces soldier and the Google security engineer, Perez was fortunate because the US judicial authorities did not file criminal charges against him; Perez did not go to jail.
During the investigation, the CFTC notified federal prosecutors in Manhattan, but the prosecutors declined to open a criminal investigation. According to informed sources, CFTC regulators have indicated a willingness to settle with Perez and have discussed terms with him. The result requires Perez to return the profits and cease similar trades thereafter.
Perez is Just the Beginning of Cleaning Up Insiders in the "Mention" Market
The reason Perez avoided jail is that prosecutors believed Perez did not constitute a criminal offense. He neither leaked important government information in advance nor caused harm to national security. As Trump said, "It's just despicable," damaging the clean image of government officials.
In March this year, the White House warned staff not to use non-public information to place bets in prediction markets. White House Spokesperson Davis Ingle stated: "The White House has strict ethical guidelines, and we expect all staff and officials to adhere to them."
But Perez is definitely not the only White House staff member profiting from insider information. Trump, who openly runs a paid group for himself, is even less qualified to comment on this teleprompter operator(Related read: $100,000 a Month, Trump Starts Selling "Alpha").
No wonder Perez couldn't resist the temptation. The "mention" market within prediction markets is indeed the category most susceptible to human manipulation. When the cost for insiders to participate is extremely low, while the potential returns are extremely high, it's no longer just a moral issue; it's a mechanism design problem. In the face of profit, even outwardly respectable, ostensibly righteous politicians cannot guarantee they will never cross that line.
The gameplay of the "mention" market involves users betting on specific words, phrases, or topics that will be mentioned in a public speech. Compared to other events (like political elections, sports events, etc.), the cheating cost for the "mention" market is extremely low. It's not limited to people like Perez who can know the speech content in advance. For the speaker themselves, cheating is as simple as saying a word, making "a word is worth a thousand pieces of gold" a concrete reality.
At the Grammy Awards ceremony in February this year, after host Trevor Noah said "Welcome back to the Grammys," he suddenly shouted "Potato." While everyone was confused, Trevor Noah continued, "If you bet on me saying that word on Polymarket, you just made a killing," and congratulated user "Noah 22." However, in reality, there was no "potato" option in the Polymarket prediction "What will be mentioned at the Grammy Awards ceremony?" and the user "noah-22" was purely fictitious.
Grammy host shouts potato at the awards ceremony
Some post-analysis suggested this was a Polymarketing marketing campaign, but it already demonstrated the speaker's ability to manipulate the "mention" market.
There's an even more direct example. In October 2025, Coinbase held its Q3 earnings call. As the call was about to end, CEO Brian Armstrong said he noticed many people were betting in prediction markets on what he would mention during this call. So he opened Polymarket and read all the words listed in the options one by one, ultimately causing the winning probability for all outcomes in that market to be 100%, ending in a draw.
The above are just two examples demonstrating a speaker's control over the "mention" market. Of course, there are certainly many more people who truly profit from it lurking beneath the surface. However, as prediction market regulation gradually deepens, perhaps all insiders in the "mention" market will be cleaned up in the future. Perez is just the beginning.
Last month, Kalshi just updated its policy, requiring users to disclose their employer. Kalshi's Head of Enforcement, Bobby DeNault, explained the reason for this move: "If you have access to certain information because of your job or employment, and you have a related legal obligation, you have a duty not to take that information for yourself or use it for personal gain." Polymarket has not yet imposed such strict disclosure requirements on users, but in the increasingly competitive and compliant prediction market track, it is believed that stricter compliance requirements from Polymarket are also coming soon.
From the special forces soldier, the Google engineer, to the White House teleprompter operator, prediction markets are gradually cleaning up insider trading. At the same time, the market is experiencing a demystification of prediction markets. Originally thought to reflect the wisdom of the crowd, they have turned out to be just ATMs for a few insiders.
Although cleaning up insider trading will make prediction markets more compliant, it also distances them further from truth and brings them closer to pure casinos.
