OpenAI fired an employee this week for using insider information to trade on prediction markets. The termination, while notable on its own, opens a window into a much larger problem: prediction markets have exploded into a $70 billion industry with virtually no enforcement infrastructure for insider trading.
And tech companies, particularly AI companies where product launches can move markets by millions of dollars, are sitting at the center of it.
The OpenAI Incident
According to TechCrunch, OpenAI terminated an employee after an internal investigation revealed they used confidential company information to trade on prediction markets including Polymarket and Kalshi. The company stated that “such trades violate its internal company policies about using confidential information for personal gain.”
OpenAI didn’t release the employee’s name. But the firing came amid mounting evidence that insider trading on OpenAI-related markets has been rampant.
Financial data platform Unusual Whales flagged 77 positions across 60 different wallets as suspected insider trades, analyzing account age, trading history, and investment patterns around OpenAI announcements.
The numbers are stark: In the 40 hours before OpenAI launched its browser, 13 brand-new wallets with zero trading history appeared on Polymarket for the first time, collectively betting $309,486 on the correct outcome.
In November 2023, two days after Sam Altman was dramatically ousted from the company, a new wallet placed a significant bet that he would return, netting over $16,000 in profits.
These aren’t isolated incidents. Unusual Whales found suspicious trading patterns around Sora’s release date, GPT-5’s launch, and multiple other OpenAI announcements.
A $70 Billion Market with Minimal Oversight
Prediction markets have grown explosively. As of February 2026, total market volume reached $127.5 billion in notional value and $69.9 billion in actual volume, with over 2.49 million unique users and more than $1 billion in open interest.
Polymarket leads with $56 billion in notional volume, followed by Kalshi at $44 billion. Monthly volume remained above $13 billion even outside major election periods.
But the regulatory framework hasn’t kept pace. Unlike traditional stock trading, where insider trading is clearly illegal under securities law, prediction markets exist in what DL News calls “a regulatory twilight zone.” The SEC doesn’t classify these platforms as securities exchanges, and the CFTC’s oversight is limited.
This matters because prediction markets increasingly cover events where insiders have valuable information: product launches, corporate announcements, personnel changes. And AI companies, with their rapid development cycles and market-moving releases, are particularly vulnerable.
Two Philosophies, Two Platforms
The two dominant prediction market platforms have taken starkly different approaches to insider trading.
Polymarket CEO Shayne Coplan has defended the practice of trading on non-public information, telling CBS News that insiders “having an edge on the market is a good thing.” His argument: allowing insiders to profit speeds up truth discovery by creating “a financial incentive for people to go and divulge the information to the market.”
While Polymarket officially bans insider trading in its U.S. rulebooks, Coplan has stated his platform can self-police through internal audits.
Kalshi has taken a different path. The CFTC-regulated platform recently disclosed its first enforcement actions, naming offenders publicly and reporting cases to federal regulators. The platform has established an independent Surveillance Advisory Committee, partnered with Solidus Labs for surveillance, and created a dedicated Head of Enforcement role.
In one case, a YouTube creator’s editor traded approximately $4,000 on video release markets with “near-perfect trading success,” resulting in a $20,000 fine and two-year suspension. In another, a California gubernatorial candidate bet on his own race, earning a five-year ban.
The CFTC Steps In
On February 25, the CFTC Enforcement Division issued its first formal advisory on prediction market misconduct, clarifying that insider trading falls under existing commodity exchange law.
The agency defined insider trading as “misappropriation of confidential information in breach of a pre-existing duty” and noted that designated contract markets must maintain audit trails, conduct surveillance, and enforce rules against prohibited practices.
But enforcement authority is limited. The CFTC can investigate and prosecute violations on regulated platforms like Kalshi, but Polymarket operates primarily outside U.S. jurisdiction through its crypto-based structure.
The AI Industry Problem
For AI companies, prediction markets pose a unique challenge. Product launches, model releases, and personnel changes can move millions of dollars on these platforms. Employees with advance knowledge face temptation that didn’t exist five years ago.
The OpenAI incident suggests companies are starting to take this seriously. But internal policies only work if companies actively monitor for violations, and prediction market trades are pseudonymous by design.
Rep. Ritchie Torres and other Democrats have co-sponsored the Public Integrity in Financial Prediction Markets Act of 2026, which would prohibit elected officials from trading on markets tied to administration policies. But broader legislation addressing corporate insiders remains absent.
What This Means
Prediction markets represent something genuinely new: a financial layer on top of information itself. When the New York Times reports a rumor, prediction markets let you bet on whether it’s true. When an AI company schedules a product announcement, markets let insiders profit on timing.
The OpenAI case is likely just the beginning. As prediction market volume continues growing and AI companies remain in the spotlight, expect more internal investigations, more enforcement actions, and eventually, clearer regulation.
Until then, every AI company employee with access to product roadmaps is sitting on information that someone, somewhere, could use to make money. Most won’t trade on it. But as the 77 suspicious wallets around OpenAI events suggest, some already have.
The Bottom Line
AI companies are learning what public companies discovered decades ago: insider information is valuable, and people will try to profit from it. The difference is that prediction markets make it easier than ever, while the rules for stopping it remain unclear.