Meta tests prediction market app with employees
Meta is testing a prediction market app, called Takes, with its employees, according to a report by The Wall Street Journal.
The app does not use real money. Participants instead wager points called “cred,” short for credibility, gaining points for correct predictions and losing them for incorrect ones. Markets on the platform include whether former Apple CEO Tim Cook will become Nvidia’s next CEO.
Meta CEO Mark Zuckerberg had previously tasked a small team with developing a prediction market product under the working name Arena.
The development comes as prediction market platforms such as Kalshi and Polymarket have grown in popularity among younger users, with just over 60% of their users under the age of 35, according to an S&P Global Market Intelligence Kagan survey.
Both Kalshi and Polymarket, however, have faced legal challenges from US state regulators over whether their sports event contracts constitute illegal gambling under state law.
Prediction market operators argue that their platforms fall under the jurisdiction of the Commodity Futures Trading Commission (CFTC) instead of state gambling regulators, a position disputed by several states, including Arkansas, Utah and Louisiana.
Takes remains in the employee-testing stage, which the WSJ described as “a step the company often takes ahead of a public release.” There is no indication of when, or whether, Meta will launch the app publicly.
Dig Deeper
The Backstory
Meta’s wager on predictions has moved from concept to internal test
Meta’s employee test of Takes marks a more concrete step into prediction markets after months of reports that the company was exploring how to turn forecasting into a consumer product. The current test is limited, uses no real money and relies on internal points called “cred,” but it puts Meta inside one of the fastest-growing and most legally contested corners of digital wagering-adjacent finance.
The company’s approach appears designed to preserve the engagement mechanics that have made prediction markets popular while avoiding the most direct regulatory trigger: cash wagering. Employees can gain or lose credibility points based on correct or incorrect predictions, with markets tied to public events or corporate speculation. That design echoes earlier reporting that Meta was building a points-based product rather than a real-money exchange, a structure that may reduce exposure to gambling laws while allowing the company to test user behavior, market design and automated question creation.
The stakes extend beyond a side project. Prediction markets have become a live competitive issue for social platforms, trading companies and sportsbook operators. Kalshi and Polymarket have drawn younger users and large trading volumes by letting participants take positions on elections, economic data, sports and other real-world outcomes. Meta’s entry, even without cash prizes, could bring prediction-style products to a scale few dedicated platforms can match.
From Kalshi talks to a standalone Meta build
Meta’s internal test follows an earlier strategic fork. The company reportedly considered buying Kalshi before deciding to build its own product, according to reporting on Meta’s proposed Kalshi acquisition talks and later Arena plans. Those talks did not result in a deal, with reported sticking points including sale terms and legal concerns.
The failed acquisition effort nevertheless showed Meta’s interest was not speculative. Kalshi held a valuable position as a federally regulated event-contract exchange and had already become one of the best-known names in prediction markets. Buying the company could have given Meta infrastructure, regulatory relationships and market liquidity. Building internally gives Meta more control but leaves it without the same real-money market framework.
Meta and Kalshi later moved to a partnership announced in March 2026 that allowed Kalshi markets to be integrated on Threads. That arrangement gave Meta a way to test audience appetite for forecasts inside its social ecosystem while keeping the regulated exchange function outside the company. Takes, by contrast, suggests Meta is also assessing a product it can operate directly, separate from the compliance burden of real-money trading.
Internal documents cited in earlier reporting indicated that Meta’s artificial intelligence systems could generate prediction questions and help determine outcomes. If used broadly, that would put AI at the center of market creation and settlement, two functions that are critical to trust. It also raises operational questions: how disputes would be handled, how ambiguous outcomes would be resolved and whether automated market selection could amplify sensitive or misleading topics.
A points model may reduce risk but not scrutiny
Meta’s reported design choices reflect the legal uncertainty around prediction markets. A points-based model differs from Kalshi and Polymarket, where users trade contracts with real financial value. The distinction matters because the central dispute in the U.S. is whether event contracts are financial instruments under federal commodity law or gambling products subject to state control.
Several states have challenged sports-related event contracts, arguing they function like illegal sports betting when offered without state gaming licenses. Operators have countered that federally overseen event contracts fall under the Commodity Futures Trading Commission, not state gambling agencies. That clash has become one of the defining legal questions for the sector.
Meta’s use of virtual points could be an attempt to sit outside both frameworks. A no-cash product may look less like gambling and less like a financial exchange. But the company’s scale could still invite attention if Takes were released publicly, particularly if points were tied to status, rewards, advertising systems or future monetization. Regulators have often focused not only on whether users can win money but also on whether a product encourages wagering behavior, targets younger users or blurs the line between games, finance and gambling.
Earlier coverage of Meta’s development of a points-based Arena app noted that the product was expected to operate separately from Facebook, Instagram and WhatsApp. That separation may help reduce reputational and compliance risks. Still, Meta’s existing audience gives any adjacent product an enormous potential funnel. The company’s platforms reach billions of users daily, meaning even a discrete app could scale quickly if promoted across Meta’s ecosystem.
Sportsbooks face a threat that analysts say is contained
Meta’s interest initially unsettled investors in companies tied to sports betting and prediction markets. DraftKings, Flutter Entertainment and Robinhood shares moved lower after reports of Arena, reflecting concern that a technology giant could pull users and liquidity away from incumbents.
Analysts have since taken a more measured view. In a June note, Jefferies analyst James Wheatcroft argued that Meta’s move was not necessarily a dire threat to online sports betting operators. His reasoning centered on licensing, product differences and the established appeal of regulated sportsbooks. If Meta avoids real money, its product may have less pull for users seeking financial upside. If it pursues real-money wagering, it would face state-by-state licensing obligations and compliance demands that are costly and slow.
That assessment frames Meta’s possible role as incremental rather than directly cannibalistic. Prediction markets can expand consumer interest in forecasting outcomes without replacing traditional sports betting, especially when contracts cover politics, corporate news, entertainment and technology. For sportsbook operators, the bigger risk may not be immediate substitution but a longer-term shift in how younger users engage with uncertain outcomes: through markets, feeds and social proof rather than conventional betting slips.
The competitive question is also about data. Meta has deep knowledge of social trends, user engagement and content distribution. Dedicated prediction platforms have trading data, market maker relationships and, in some cases, regulatory approvals. Sportsbooks have pricing models, customer accounts and gaming licenses. The company that can combine engagement, trust and compliance most effectively may shape the next phase of the category.
Consumer protection concerns are widening
The growth of prediction markets is unfolding against a broader debate over online gambling safeguards. A report from the Center for Addiction Science, Policy and Research found that states with legalized online sports betting or igaming generally scored poorly on consumer protections, while states without online gambling received higher grades. The group’s state-by-state assessment of online betting protections highlighted issues such as loss limits, prop and micro-bet restrictions, operator intervention and efforts to block prediction market platforms.
That context matters for Meta because a prediction product built around points can still normalize wagering-like behavior. The policy concern is not limited to financial loss. It includes frequency of engagement, compulsive use, social competition and the gamification of real-world uncertainty. A product that rewards users for being right about news events could be framed as educational or social, but it could also train behaviors similar to betting if poorly designed.
The age profile of prediction market users adds to the sensitivity. Surveys have found a majority of Kalshi and Polymarket users are under 35, a demographic also central to social media platforms. Regulators and public health advocates are likely to watch whether prediction products use leaderboards, streaks, notifications and social sharing to drive repeat participation.
For Meta, the safer initial path is an internal test with no cash and uncertain launch plans. But the internal rollout still signals intent. The company is studying how to convert forecasting into a repeat-use product at a moment when prediction markets are growing quickly, courts and regulators are still defining the rules and gambling policy is under renewed scrutiny. Whether Takes remains an employee experiment or becomes a public app, it reflects a larger convergence of social media, financial speculation and betting culture that regulators have not yet fully caught up with.









