Meta Is Testing a Prediction-Market App Called Takes, Where Employees Bet Cred Points Instead of Real Money

Meta is quietly building a prediction market of its own. According to the Wall Street Journal, citing people familiar with the matter, Meta employees have begun testing an in-development prediction-market app called Takes, betting on the outcomes of questions and events ahead of any public release — the classic dogfooding step Meta applies to most products before launch.

The mechanics distinguish it from the commercial leaders. Employees wager "cred" — short for credibility — rather than real money: correct predictions earn cred points, wrong ones cost them. That is a fundamental contrast with Kalshi and Polymarket, both of which let users trade real dollars and both of which face lawsuits from multiple US states alleging they operate illegal gambling platforms; the companies maintain they are federally regulated under the CFTC.

The internal questions on Takes track the news cycle closely. One of the markets live on the app Thursday asked whether Apple's Tim Cook would become Nvidia's next chief executive — the question consuming Silicon Valley this week — and most employees bet he would not. Whether Meta ships the app, or ships it in this form, remains open: the company declined to comment, and internal test products die more often than they launch.

The project has been running for a while. The Journal reports that the app was known internally as Arena earlier this year, after CEO Mark Zuckerberg instructed a team to develop a prediction market. In a recording of a spring internal meeting, Zuckerberg was heard saying he and chief product officer Chris Cox had discussed building up to 50 new apps — a scope that goes well beyond one experiment.

The commercial logic points at a specific audience. Prediction markets have surged in popularity among Gen Z users — just over 60 percent of Kalshi and Polymarket users are under 35, according to a recent S&P Global Market Intelligence Kagan survey — exactly the demographic Meta has been struggling to hold inside its own apps. An engagement product where users stake points on outcomes, share their calls, and argue about the results is, structurally, a social feed with better retention hooks.

If Meta does ship Takes publicly, it will land in a legal gray zone even without real money: the credibility-point model avoids most gambling statutes, but the underlying format — and any later drift toward paid markets — would run into the same state-level fights that Kalshi and Polymarket are fighting now. The safer read is that Meta is prototyping the format while the regulatory map settles. The company has launched a hundred products that never left dogfooding; what matters is that one of the three or four companies that define consumer internet behavior now considers prediction markets a product surface worth testing.