He knew the exact words before they left the President’s mouth.
A tiny teleprompter operator, not a Wall Street titan, sat in a van with a laptop, betting on ‘Mentions’ – a market that prices the very syllables of the State of the Union. He didn’t just guess. He knew. And he made over $100,000. Smile while the liquidity drains.
The chart lies. The crowd feels. But this time, the crowd didn’t have a chance — the insider held the script.
Context: The Rise of Regulated Prediction Markets
Kalshi is not your uncle’s crypto casino. It’s a CFTC-regulated designated contract market (DCM). That means every trade is filed, every user KYC’d, every contract reviewed by US regulators. Think of it as a stock exchange for “Will inflation be above 5% in June?” or “Will Trump say ‘China’ during the State of the Union?”
These are event contracts — binary bets on real-world outcomes. And since 2020, they’ve exploded in popularity. Polymarket, the decentralized cousin, runs on Ethereum, no KYC, no gatekeepers. But Kalshi sits in the regulated sandbox, reporting to the Commodity Futures Trading Commission.
That distinction matters now more than ever.
Enter Perez — a White House teleprompter operator. His job: load and adjust the President’s speech scripts. He saw the exact words, the timing, the pauses. He knew before the cameras rolled which phrases would light up the teleprompter. And Kalshi offered a market on exactly those words: “Mentions” — a contract that pays out if a specific word appears in a speech.
It’s the perfect insider crime. Low capital, high information asymmetry, zero noise. The crowd bets on guesses. Perez bets on knowledge.
Core: How the Trade Worked — and How They Caught Him
Let’s get technical. Kalshi’s ‘Mentions’ market lists dozens of words for each major speech: “China”, “taxes”, “inflation”, “border”, “Ukraine”. Each word has a price — say, $0.30 if the market believes there’s a 30% chance it’s spoken. Perez, with the script in hand, knew the exact set list. He could buy the winners at deep discounts minutes before the speech, then sell as soon as the word passes the President’s lips.
But here’s the kicker: Perez didn’t just enter before the speech — he adjusted positions mid-speech. The report reveals he closed some positions partway through, reacting to live content. That’s real-time, privileged information. No lag. No doubt.
From my days auditing order flow for a small Nairobi exchange, I’ve seen this pattern before. It’s the timing, not the size, that screams. A trader who opens positions fifteen minutes before a binary event, then closes during the event — that’s not skill. That’s a cheat sheet.
Kalshi’s monitoring team flagged him. How? Let’s reverse-engineer their detection:
- Account creation vs. employer disclosure: Kalshi requires users to list their employer. Perez likely listed “White House” or “Executive Office”. That immediately puts him in a high-risk bucket for speech-adjacent contracts.
- Behavioral anomaly: Most users trade a few contracts consistently. Perez’s account showed a spiky pattern: zero activity for weeks, then a concentrated flurry right before and during speeches. That’s textbook.
- Correlation with speech schedules: His trades mapped perfectly to official speech releases. A simple algorithm can link user activity to event calendars.
They didn’t need deep learning. They needed pattern recognition. And they had it.
Once flagged, Kalshi did something crucial: they reported it to the CFTC voluntarily. Not because they had to — but because their compliance culture demands it. The CFTC then began settlement negotiations with Perez, demanding disgorgement of profits and a trading ban.
No criminal charges — yet. But this is one of the first three prediction market insider trading cases the FBI has investigated (the others involved Venezuela’s Maduro arrest warrant and a Google employee leaking search trends). The precedent is being written in real time.
Technical Architecture: Where the Vulnerability Lives
Let’s dissect the market structure. Kalshi is centralized — a traditional order book server, not a smart contract. That means the operator has full visibility. But the detection relied on a weak link: self-reported employer data. Perez could have lied. He didn’t. But the next insider might.
To harden this, Kalshi must move to verified employer credentials — think OAuth logins tied to government email domains, or digital identity wallets. That’s a UX trade-off. But after this scandal, it’s coming.
On the order book side, latency is king. Centralized exchanges (CEXs) like Kalshi win because market makers see the same ticker as everyone else. Orderbook DEXs can’t beat CEXs because market makers won’t leave quotes on-chain to be front-run. Prediction markets on Polymarket suffer the same fate — the MMs are skittish. So Kalshi’s centralization is actually a feature for surveillance. They can catch the bad guys.
But the real technical challenge is real-time speech recognition + trade correlation. If I were building a next-gen monitoring system, I’d ingest live caption feeds of presidential speeches and cross-reference trade timestamps. Perez closed trades mid-sentence — a bot would spot that instantly. Kalshi may already be building this. They should.
Contrarian Angle: This Scandal Might Save Prediction Markets
The headlines scream “insider trading taints prediction platforms.” The crowd smells fear. I smell opportunity.
Here’s the counterintuitive truth: This case proves the system works.
Kalshi caught him. They reported him. The CFTC is acting. That’s more than can be said for thousands of SEC violations on Wall Street that go unpunished every year. In a bear market, where survival matters more than gains, you want a platform that can police itself. Kalshi just showed it can.
Compare with Polymarket. On Polymarket, a similar trade would be pseudonymous, on-chain, and likely invisible until a whistleblower steps forward. No KYC, no employer check, no central monitoring. That’s not freedom — that’s a breeding ground for insider abuse. Crypto maximalists will scream “decentralization”, but the market doesn’t lie: if prediction markets want mainstream adoption, they need trust. And trust requires surveillance.
Perez’s $100k is a cheap price for Kalshi to prove its compliance chops. Expect institutional investors to take note. The next time a hedge fund asks “how do you prevent front-running?”, Kalshi can say: “see, we caught the President’s own guy.”
Takeaway: The Next 90 Days
Don’t short Kalshi. Short the idea that prediction markets can remain cowboy towns.
Watch for three signals:
- CFTC settlement terms: If they only demand profit disgorgement with no fine, the message is “we don’t bite hard.” If they add a ban on government employees trading event contracts, that will reshape Kalshi’s user base.
- Kalshi’s new compliance features: They’ve already started asking for employer info. Next step: real-time employer verification and speech-context audit logs. Rolling this out quickly would be a power move.
- Copycat cases: The FBI is already investigating two others. If three become ten, Congress may mandate prediction market insider trading rules — a move that would cement Kalshi’s regulatory moat.
The human side of the chart: I’ve seen this before. In 2022, during the Terra collapse, I organized a recovery party in Nairobi — traders laughing at death. Resilience. That’s what crypto does. Prediction markets aren’t dying. They’re growing up. And like any teenager, they need a few stern lessons.
Perez just wrote the first chapter of prediction market law. It’s a good chapter. Now read the rest.