1. Meta's prediction market app reaches its own staff
Meta Platforms has opened Takes, its in-house prediction market app, to employees for testing, with a points system called cred replacing real money, people familiar with the matter said in a report published November 13. Staff use cred to back their forecasts on future events, from corporate leadership changes to television quiz results, and no cash changes hands.
The arrangement sets the Meta prediction market app apart from the two best-known names in the sector, Kalshi and Polymarket. On both of those platforms, users put actual dollars behind their views on how real-world events will turn out, and the value of a position rises or falls with the perceived likelihood of the outcome. Takes keeps the forecasting mechanics but strips out the money.
The test is limited to Meta's workforce. The company has not announced the product, has not set any timeline for a public release and declined to comment on the app. Even so, the internal rollout is the clearest sign yet that one of the world's largest social media companies is experimenting seriously with a category of product that has become one of the most contested corners of the financial and technology industries.
2. How does cred work?
Cred is a points balance that rewards accuracy. The name is short for credibility, and the logic follows from it: employees who call an outcome correctly see their cred grow, while those who back the wrong answer lose some of what they staked. Over time, a user's balance becomes a running scorecard of how good their judgment has been.
Because cred cannot be converted into money, the stakes are reputational rather than financial. A participant who builds a large balance has demonstrated a track record, but has not earned anything that can be withdrawn or spent. That is the central distinction between Takes and the cash-based platforms it resembles, and it is a distinction that carries considerable legal weight in the United States at the moment.
Points-based forecasting is not a new idea in itself. Companies and research groups have long used internal markets and scoring systems to gather the collective judgment of employees or volunteers, on the theory that people who are rewarded for accuracy tend to think harder about their predictions and that a crowd's aggregated view can be more reliable than any single expert. What makes the Meta experiment notable is the scale of the company behind it and the timing, as the money-based version of the same concept faces mounting legal pressure.
For Meta, a points model also lowers the barrier to participation. Employees can weigh in on any question that interests them without putting personal savings at risk, which encourages broad involvement and produces a larger pool of forecasts for the company to study as it evaluates whether the concept has wider appeal.
3. From an app called Arena to Takes
Work on the product goes back to at least the summer. Meta Chief Executive Mark Zuckerberg reportedly instructed a team to build a standalone prediction market app, which at that stage carried the internal name Arena. The project has since surfaced among employees under the Takes name.
Zuckerberg has not spoken about the app in public. A recording of an employee meeting held in the spring, however, captures him and Meta's chief product officer, Chris Cox, discussing as many as 50 new apps that the company was weighing. The prediction market concept emerged from that broader push to test a large slate of new standalone products.
Zuckerberg's interest in prediction markets had already drawn attention earlier in the year. In June, the plans that were described at the time did not rule out the possibility of real-money trading at some later stage. The version now in employees' hands uses points only, but nothing that has been disclosed so far settles what any eventual public product would look like.
4. What employees are betting cred on
The questions on Takes range from the weighty to the light-hearted. On November 12, one market asked whether Tim Cook, formerly Apple's chief executive, would be named the next chief executive of Nvidia. Most of the employees who took part bet that he would not.
Another market on the app asked participants to forecast which contestant would win that Friday's episode of the long-running television quiz show Jeopardy!. The pairing of a question about leadership at one of the most valuable companies in the world with a question about a game show illustrates the breadth of topics the app is designed to handle.
That breadth mirrors the commercial platforms. Kalshi and Polymarket list contracts on politics, economics, sports, entertainment, weather and corporate events, among other subjects. Each contract typically resolves to a yes or no answer, and the trading price before resolution is often read as the market's estimate of the probability that the event will occur. Takes appears to borrow that format while replacing the dollars with cred.
5. Dogfooding: a familiar step that often goes nowhere
Testing an unreleased product on employees first is standard practice at Meta, where the approach is known as dogfooding. The term comes from the idea of a company using its own products before asking customers to do so, and it gives engineers and product managers a chance to see how real people interact with a feature before it reaches the public.
Most products that go through this stage never ship. Many are shelved after internal testing reveals weak engagement, technical problems or strategic doubts. Those that do make it out often arrive with a different name, a different design or a narrower set of features than the version employees first saw. The move from Arena to Takes is itself an example of how much a product can change while it is still in development.
For that reason, the internal launch of Takes is best understood as an experiment rather than a commitment. No public release date has been reported, and Meta has offered no comment. The company has a long history of testing standalone apps, and the presence of a prediction market among those tests says more about the direction of its curiosity than about any finished plan.
6. Why Kalshi and Polymarket are fighting state gambling suits
The decision to use points rather than money lands in the middle of an unresolved legal battle. Several US states have sued Kalshi and Polymarket, arguing that the platforms are running illegal gambling operations. In the states' view, a wager on the outcome of a sporting event, an election or an awards show is a bet, whatever label the platform attaches to it, and should be subject to state gambling laws and licensing.
The companies reject that characterization. They argue that their contracts are financial instruments overseen by the Commodity Futures Trading Commission (CFTC), the federal agency that regulates derivatives markets in the United States, including futures and options. If federal oversight applies, the platforms contend, state gambling rules should not.
The outcome of that dispute matters well beyond the two companies involved. It will help determine whether event contracts can be offered nationwide under a single federal framework, or whether operators must navigate a patchwork of state rules that may prohibit them outright. Until the question is settled, any company considering a money-based prediction market in the United States faces significant legal uncertainty.
7. Points keep Takes clear of the wagering question
By keeping money out of the employee version of Takes, Meta has built a prediction market that likely sidesteps the wagering issue at the heart of the state lawsuits. Gambling laws generally hinge on the presence of something of value being staked in the hope of winning something of value. A points balance that cannot be cashed out is a very different proposition from a dollar-denominated contract that can be sold or redeemed.
That design choice allows Meta to study how people use a prediction market, what kinds of questions attract participation and how accurate the crowd's forecasts turn out to be, without exposing itself to the legal fight that Kalshi and Polymarket are waging. It also lets the company gather data from a large and engaged internal user base before deciding whether a public product is worth pursuing.
Whether that remains the approach is an open question. The earlier plans described in June left room for real-money trading in the future, and the current points-only design applies to an internal test. Any move toward cash would bring Meta directly into the regulatory and legal debate that the points system currently avoids.
8. What comes next for the Meta prediction market app
For now, Takes remains an employee-only experiment with no announced release date. Meta has not confirmed the project publicly, and Zuckerberg has not addressed it. The company's history suggests that the product could be expanded, reshaped, renamed or quietly dropped depending on how the internal test goes.
The broader prediction market sector, meanwhile, will continue to be shaped by the state cases against Kalshi and Polymarket and by the question of how far CFTC oversight extends. Those cases will determine the legal ground any large technology company would have to stand on if it ever chose to let users stake real money on future events.
Until then, Meta's employees are placing their bets in cred, on questions that run from who might lead Nvidia to who will win a quiz show, and the company is watching to see what the experiment reveals.
