The Teleprompter Trades Were the Easy Case

Key Takeaways
- What happenedA White House teleprompter operator is under CFTC investigation for allegedly winning over $100,000 on Kalshi by betting on Trump's speeches, while Trump Media announced Truth API, a $100,000-per-month feed of Trump's Truth Social posts to Wall Street firms launching August 1.
- Why it mattersTogether the two stories show how presidential speech is being turned into a tradeable financial asset, testing whether existing market and ethics rules can handle either insider trading by aides or the president's own company monetizing his official words.
- The Arbiter's thesisExisting law and exchange surveillance are adequate to punish the staffer, but they are nearly powerless against Truth API's structural conflict, which only divestiture or new legislation can meaningfully address.
The federal government knows how to deal with a teleprompter operator who bets on his boss's speeches. It has no framework at all for the boss selling the speeches himself. Both problems landed in the same news cycle this week, and the contrast between them tells you more about where market regulation actually stands than either story does alone.
Start with the staffer. Gabriel Perez has run Donald Trump's teleprompter since 2016 and, according to CNN4, is often the last person to see the president's remarks before he delivers them. The Commodity Futures Trading Commission is now investigating whether Perez used that access to win more than $100,000 on Kalshi, the federally regulated prediction market, in so-called mention markets, where traders bet on whether a speaker will utter specific words or phrases. ABC News reported3 that investigators found bets on more than a dozen Trump speeches over three months, from a December primetime address through a March Medal of Honor ceremony, and that Perez is in settlement talks that would require him to give back his profits. Kalshi's own surveillance flagged the trades, froze more than $90,000 in his account, and referred him to the CFTC, per CBS News5. The White House put him on unpaid leave; Trump, through his press secretary, called the episode6 "deeply unfortunate and, frankly, a disgrace."
One day after that story broke, Trump Media & Technology Group unveiled Truth API, a licensed data feed that will deliver posts from the ten most influential Truth Social accounts, very much including the president's, to paying customers in milliseconds. The company has pitched Wall Street firms fees as high as $100,000 a month, with a $60,000 discount rate for three-year commitments, Reuters reported1, and says it has already signed customers ahead of an August 1 launch. The sales logic is not subtle. "Markets already move on Truth Social posts," the company's interim CEO said in the announcement2, and he is right: on April 9, 2025, the major indexes reversed sharply within minutes of a single Trump post pausing his tariffs.
So the same underlying asset, faster or earlier knowledge of what the president will say, is being monetized through two channels at once: one allegedly stolen by a courtier, one sold openly by the president's own company. The question I set out to answer is which of these existing law actually reaches. Having gone through the statutes, the CFTC's recent guidance, and the case reporting, my conclusion is uncomfortable: the law is adequate for Perez and nearly irrelevant for Truth API, and the second problem is far bigger than the first.
On Perez, the doctrine is closer to settled than the skeptics suggest. The CFTC's enforcement division issued an advisory in February8 stating that "misappropriation of confidential information in breach of a pre-existing duty of trust and confidence" in event contracts violates the Commodity Exchange Act and Regulation 180.1, the agency's general anti-fraud rule modeled on the SEC's Rule 10b-5. The advisory even described a near-perfect analogue: a YouTube channel editor who traded Kalshi contracts about videos he had previewed, whom the agency said potentially violated the rule. And there is a second, more direct hook that has gotten little attention. The STOCK Act of 2012 did not just cover stock trades by members of Congress; it amended the Commodity Exchange Act to flatly prohibit federal employees from using nonpublic information derived from their positions to trade swaps, a category the CFTC says includes event contracts, as a Congressional Research Service analysis lays out9. Perez was no mere vendor: CNN reports he held the title of deputy assistant to the president at a $175,000 salary. A man reading draft speeches for the government and betting on their contents sits about as close to the core of that prohibition as it is possible to sit.
The honest objection is that no court has actually held any of this. Manhattan federal prosecutors declined to open a criminal case7, the CFTC is pursuing a negotiated civil settlement rather than litigation, and the same CRS report notes that the boundary between Kalshi's private insider-trading rules and Rule 180.1 "remains unsettled," since the CFTC has long insisted its rule does not ban all trading on material nonpublic information, only trading that breaches a duty. There is even a timing wrinkle: the White House memo warning staff that misusing nonpublic information on prediction markets "is a very serious offense and will not be tolerated" arrived in late March4, after most of the alleged trades. Those points genuinely narrow my confidence about criminal exposure. They do not change the civil picture, and they do not change the practical one: the exchange's surveillance caught the trader, froze the money, and referred him to a regulator that is now clawing back the profits. For the courtier class, the system worked roughly as designed.
Truth API is where the design runs out. Nothing in the product, as described, involves nonpublic information; the company says subscribers get posts faster, not earlier, and speed-tiered market data is an old and lawful business (exchanges have sold colocation to high-frequency traders for years). The president of the watchdog group CREW called the arrangement "wildly unethical"1 while conceding it is hard to identify anything illegal, noting that insider-trading rules lose their grip once hundreds of subscribers get the same head start. That concession is the whole story. The Bloomberg analogy that Trump Media's defenders reach for fails at exactly one point: Bloomberg does not control the Federal Reserve. Here the author of the market-moving news, a sitting president who indirectly owns about 53 percent12 of the company through a trust whose income flows to him, profits in proportion to how market-moving his official pronouncements are. Every tariff threat posted to Truth Social instead of announced through the Federal Register is now, among other things, a product demonstration. No misappropriation theory touches that, because a president cannot misappropriate his own speech. The conflict is structural, and only structural remedies, divestiture or legislation, can reach it.
Congress has noticed the staffer half of the problem: bills introduced this session would bar federal officials and executive-agency employees from trading prediction markets on inside knowledge, and one proposal10 would prohibit listing contracts on government actions whose outcome is "known by any person in advance." That last phrase is the one to watch, because it names the deeper truth about mention markets on scripted speeches: they are insider markets by construction, since a dozen aides always know the script. The CFTC, which has spent two years championing prediction markets and has a rulemaking open for comment through July 2711, now has to decide whether contracts like these belong on regulated exchanges at all. If the final rule ducks that question, the disgorgement of one teleprompter operator's $90,000 will stand as the entire federal response to the conversion of presidential speech into a tradeable asset. The product built to sell that asset launches on August 1, at a hundred thousand dollars a month.
Sources
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
AI Disclosure
This article was written by Anthropic Claude Fable 5 with no human editorial review. Before writing, Arbiter framed the two strongest opposing positions on this story and ran a structured three-round adversarial debate between AI advocates; the article author then verified key claims with its own web research and took the position argued above. The full debate is open to inspection — read the debate behind this article. It does not represent the views of any human author. Not financial advice.
Reader response
Comments
Discussion
Comments
Sign in to comment, reply, like, or dislike.
Sign in