Goldman Sachs Bans Staff from Prediction Market Trades as Insider-Trading Risk Grows Across Corporate America

The corporate world is starting to show what an internal response to prediction market insider-trading risk looks like — and the gap between the few who have acted and the many who have not is stark.

Goldman Acts. Almost Everyone Else Has Not.

Goldman Sachs has told employees they cannot trade prediction market contracts tied to bank-specific events, elections, financial markets, macroeconomic data releases, or geopolitics, according to people familiar with the policy as reported by CNBC. A Goldman spokesperson declined to detail the directive publicly but confirmed the bank prohibits using material, nonpublic information to trade in any market.

CNBC contacted 50 publicly traded and privately held companies, all of which have prediction market contracts referencing details about their own businesses. Only three said they have policies governing employee trading on these platforms. Another two said it was something they were actively reviewing. A total of 36 companies did not respond to CNBC's inquiries, and another seven declined to comment.

The Case That Changed the Math

In May, the Commodity Futures Trading Commission and the Department of Justice charged Google employee Michele Spagnuolo with trading Polymarket contracts using inside knowledge of Google's "Year in Search" lists. Trading under the handle "AlphaRaccoon," Spagnuolo allegedly collected roughly $1.2 million in profit, according to the CFTC's complaint. The charges are allegations; no conviction has been entered.

This was the first prediction market insider-trading action involving a private-sector company, not a government official or political operative. The CFTC and DOJ moving against a rank-and-file tech employee signals the agencies are watching these platforms, not just theorizing about them.

Why This Is Harder to Police Than Stock Trading

Traditional securities insider trading has a well-worn enforcement map: you know which stocks a company's employees might have an edge on, and you watch for unusual options activity ahead of earnings or M&A. Prediction markets blow that map up.

Karen Woody, a law professor at Washington and Lee University, told CNBC the problem plainly. "All these different questions that you're able to bet on… it makes it really hard to kind of play whack-a-mole in terms of where people are using the information they've obtained confidentially."

A Google employee with access to internal hiring data could trade contracts on what the company's headcount will be. An employee on the Gemini AI team could trade contracts on when the next version ships. None of these are obvious stock trades, and they may not have been covered by existing compliance policies until recently.

The Compliance Gap Is Real

David Oliwenstein, a partner and securities enforcement practice lead at Pillsbury, told CNBC his firm is fielding steady demand. "We are getting constant questions from clients, particularly among regulated entity clients, about what the regulator expectations are, what the risks are, where the areas of potential liability are."

Lawyers are telling clients the window to act proactively is still open. Oliwenstein and others warn that as more cases are prosecuted, regulators will expect companies to have had training and written policies in place. A company caught flat-footed after the Spagnuolo case will have a harder time arguing it had no reason to see the risk coming.

Marissa Mastroianni, an employment law attorney at Cole Schotz, noted: "Right now, training is not necessarily the gold standard, just because it is new."

The Counterargument Worth Hearing

Some legal observers and company representatives argue that broad directives banning insider trading already inherently apply to prediction markets. A person familiar with OpenAI's employee policies said that the company's blanket insider trading policy makes clear that staff cannot use material, nonpublic information in any way.

But Tiffany Magri, a regulatory advisor at compliance technology company Smarsh, said companies benefit from explicitly mentioning prediction markets in their policies. "The question is no longer whether exchanges can detect suspicious trades," she said. "It's whether employers have established clear expectations around when employees should be prohibited from participating."

Blanket prohibitions, like Goldman's, have a real cost to market quality that policymakers will eventually have to weigh.

What Comes Next

The CFTC has jurisdiction over event contracts through its oversight of designated contract markets, and the DOJ's willingness to bring criminal charges in the Spagnuolo case suggests federal prosecutors view prediction market fraud as a priority, not a novelty. Whether the CFTC issues formal guidance telling firms what compliance expectations look like — rather than waiting to litigate it case by case — is the open question that will determine how fast the companies that have not yet responded move off the sideline.