Meta Moves Into Forecasting
Mark Zuckerberg has directed Meta to develop its own prediction market platform, according to reports – a move that plants the social media giant squarely in territory already occupied by dedicated forecasting services.

What a Prediction Market Actually Is, and Why Meta Wants One
Prediction markets let users stake money or points on the outcomes of real-world events – elections, sports results, economic indicators, corporate decisions. The mechanics are straightforward: if you believe something will happen, you buy a position. If you’re right, you profit. The collective pricing of those positions produces a crowd-sourced probability estimate that has historically outperformed traditional polling and expert analysis in several documented cases.
Platforms like Polymarket and Kalshi have spent years building audiences around this concept, and both saw significant traffic spikes during the 2024 U.S. presidential election cycle. Polymarket alone processed hundreds of millions of dollars in election-related volume. That kind of engagement – opinionated, sticky, return-visit-driven – is exactly the kind of behavior Meta’s platforms have struggled to sustain as younger users drift toward short-form video and away from text-based debate.
Zuckerberg’s reported directive suggests Meta sees prediction markets not just as a product feature, but as a mechanism for re-anchoring user attention around live events. Rather than watching a news cycle scroll past in a feed, users with a financial or point-based stake in an outcome have a reason to keep checking back. That retention logic maps cleanly onto Meta’s broader interest in keeping users inside its apps longer – particularly on Facebook, where the average age of the active user base continues to climb.
There is also a data angle. Prediction markets generate structured, timestamped expressions of user belief about specific topics. That behavioral signal – who bets what, when, and on which events – is a richer dataset than a like or a share. For a company that has built its advertising business on behavioral inference, the informational byproduct of running a prediction market is not a small consideration.

The Complications Meta Will Have to Navigate
Building a prediction market is not the same as launching a new feed format or a Stories clone. The regulatory environment is complicated. In the United States, real-money prediction markets that touch on elections operate in a legal gray zone that the Commodity Futures Trading Commission has spent years trying to define. Kalshi fought a prolonged legal battle with the CFTC before being permitted to offer U.S. election contracts. Polymarket, which is based offshore and technically restricted for U.S. users, became the most-discussed platform anyway – a fact that illustrates how porous those restrictions are in practice.
Meta’s scale creates a specific version of this problem. A small platform running election markets can be dismissed as a novelty. Meta, with billions of active users across Facebook, Instagram, and WhatsApp, would be running one of the largest public forecasting systems ever built the moment it launched. That draw regulatory attention almost by default, regardless of whether the markets involve real money or virtual tokens.
The virtual-token route is the most likely starting point, precisely because it sidesteps direct gambling regulation. Platforms including X (formerly Twitter) have experimented with token-based forecasting tied to trending topics without triggering the same regulatory scrutiny as real-money markets. Meta could plausibly launch something similar – users predict outcomes, earn or lose points, climb leaderboards – without ever processing a financial transaction in the regulatory sense.
But a points-only system creates its own tension. Part of what makes prediction markets function as forecasting tools is that real money creates real incentives to be accurate. When users are betting virtual currency with no cash value, the quality of the crowd-sourced signal degrades. Meta would be building the interface of a prediction market without the epistemic engine that makes those markets interesting. Whether that matters depends on what Meta actually wants: a forecasting tool, or a stickiness mechanism dressed up as one.
Meta has also been navigating a complicated public relationship with political content. The company spent several years actively reducing the distribution of political posts on Facebook and Instagram, citing user feedback that political content felt exhausting and divisive. Prediction markets centered on elections and political events would reverse that direction entirely – pulling political forecasting back to the center of the product experience. Meta has previously partnered with law enforcement agencies to dismantle coordinated inauthentic behavior at scale, and a live prediction market tied to elections would create new surfaces for that kind of manipulation.

Where This Leaves the Existing Players
For Polymarket and Kalshi, Meta’s reported entry is the kind of news that is hard to read cleanly. On one hand, a major platform validating the prediction market concept accelerates mainstream awareness in a way that years of press coverage have not managed. On the other hand, Meta has the distribution, the existing user accounts, and the engineering resources to build a competing product that reaches more people in its first week than either platform has reached in its entire history.
Zuckerberg’s track record with adjacent-platform acquisitions and copies – Instagram, WhatsApp, the relentless attempts to replicate TikTok’s format – suggests that when Meta decides a product category is worth entering, it commits resources that independent startups cannot match. Whether the company builds this internally, acquires a smaller player, or partners with an existing market operator is still unknown. What is known is that one directive from Zuckerberg is now sitting somewhere inside Meta’s product organization, waiting to become a roadmap item – and the platforms that built this space without Meta’s backing are now watching to see which version of the company shows up.








