Author: Nancy, PANews

Meta has once again been an early mover but a late arrival.

After a six-year hiatus, the tech giant Meta is re-entering the prediction market arena. As early as 2020, Meta quietly launched an experimental prediction app called Forecast, attempting to explore the commercial viability of collective intelligence. However, the product failed to make significant waves and eventually faded away quietly.

Today, prediction markets have grown into a red-hot new frontier, with players like Polymarket and Kalshi pushing this sector to a scale of tens of billions of dollars. Now, Meta is making another move, with Mark Zuckerberg personally driving the internal development of a prediction market app called Arena, aiming to leverage its massive user ecosystem and traffic advantages to claim a share of this feast.

Zuckerberg personally leads the charge, with 3.5 billion daily active users as Meta's biggest bargaining chip

According to The New York Times, Zuckerberg recently assigned a small internal team at Meta to develop a smartphone app similar to Polymarket and Kalshi, internally codenamed "Arena."

Sources say Arena will allow users to make predictions on future events and will operate independently of Meta's existing social product ecosystem, including Facebook, Instagram, WhatsApp, and Messenger.

Unlike existing prediction market platforms, the initial version of Arena will not involve real-money betting mechanisms but is more likely to adopt a game-like points system.

This design stems from multiple practical considerations.

First, real-money prediction markets face complex gambling regulations, compliance reviews, and potential litigation risks in the U.S. and globally. Adopting a points system helps position Arena as a social entertainment prediction tool, thereby reducing legal and compliance pressures, alleviating external concerns about addiction, market manipulation, and money laundering risks, and reserving space for global expansion.

Second, as a product still in the early validation stage, a low-barrier points mechanism allows Meta to quickly test prediction mechanisms and social interaction models at lower legal and operational costs, continuing its consistent rapid trial-and-error strategy.

Furthermore, the design without monetary thresholds also helps lower the participation barrier, expand user coverage, and accelerate the formation of scaled participation and community interaction.

Although the project is still in the experimental stage, Zuckerberg has personally pushed it forward and listed it as one of the internal priority projects.

Compared to most new entrants, Meta plans to leverage its vast social network user base to drive traffic and growth for Arena. According to Meta's Q1 2026 earnings call disclosure, its social platforms reached 3.56 billion daily active people (DAP) that quarter, which is seen as a key foundation for its rapid entry into the prediction market.

However, Meta insiders also revealed that Arena is still in early development, and whether it will ultimately be officially released remains undecided.

Failed trial six years ago, Meta battles prediction markets again

In fact, this is not Meta's first foray into the prediction market track.

As early as 2020, Meta's New Product Experimentation (NPE) team launched the crowdsourced prediction app Forecast. Users could use a points system to predict and discuss future events. At the time, NPE was positioned as Meta's innovation incubator, focused on rapidly testing various new social product formats. Forecast was one of the apps NPE tested then.

The birth of Forecast stemmed, to some extent, from the multiple pressures Facebook faced at the time. Back then, Facebook was grappling with declining quality of social discussions, the spread of fake news and conspiracy theories, and intensifying polarization in comment sections. Therefore, Meta hoped Forecast would guide users to express opinions based on evidence rather than emotional outbursts.

Simultaneously, Meta was also trying to verify whether collective public predictions could outperform expert judgment in specific areas (such as pandemics, economics, and public events). Early Forecast topics once focused on pandemic-related predictions to test whether collective intelligence could evolve into a new type of information source.

Rebecca Kossnick, then head of NPE, once stated: "We believe a community built around predictions can not only pool crowdsourced wisdom but also potentially foster healthier online conversations."

However, due to vague product positioning, insufficient user incentives, limited mechanism design, gradually shrinking team resources, and the controversial nature of pandemic-related topics, Forecast struggled to achieve stable growth and was eventually shut down two years later.

Similar stories of missing the wave are not uncommon in Meta's product history. Take stablecoins, for example. In June 2019, Meta grandly announced its stablecoin plan Libra, but subsequently faced global regulatory pressure, political resistance, and the withdrawal of multiple partners. The project was forced to rebrand as Diem and continuously scaled down, ultimately being sold to Silvergate Bank in 2022.

But this year, when Meta returned to the stablecoin track with a more cautious stance, the market had already "changed dramatically." Tether and Circle had firmly seized dominant shares, while more and more traditional financial giants entered the fray, causing the competitive landscape to solidify. Meta had no choice but to abandon issuing its own stablecoin and instead integrate third-party stablecoin payment systems, shifting its focus to leveraging its advantages in user experience, distribution capabilities, and platform scenarios.

Prediction market enters oligopoly landscape, making it harder for latecomers to grab a share

Six years later, Meta's return to the prediction market table is no easy feat. What was once a niche experiment has now evolved into a hotly contested battlefield where giants and emerging players fiercely compete.

According to Dune data, as of June 24, the cumulative number of unique addresses participating in prediction markets has exceeded 3.77 million, with cumulative nominal trading volume surpassing $259.3 billion. Looking at monthly performance, nominal trading volume in June alone reached $33.6 billion, an increase of about 15.6 times compared to the same period last year; active participating addresses exceeded 740,000, nearly 2.9 times the figure from last year; monthly fee revenue exceeded $220 million, a year-on-year increase of about 27.5 times.

Despite the market's rapid expansion, the share has already been carved up by a few leading players, with Kalshi and Polymarket holding absolute dominance.

Dune data shows that as of June 24, in terms of nominal trading volume, Kalshi accounts for 62.8% and Polymarket 23.7%; in terms of fee capture, Kalshi holds 77.1% and Polymarket approximately 21.1%. Meanwhile, the valuations of both have rapidly climbed to around $22 billion and $15 billion respectively.

Moreover, the prediction market is welcoming more entrants, including Charles Schwab, DraftKings, Robinhood, Coinbase, and Webull, all of which have announced or launched related business layouts.

Beyond competitive pressure, prediction markets are facing increasingly severe regulatory and legal challenges. Many countries globally have imposed varying degrees of restrictions or bans on prediction markets. Recently, U.S. regulators also issued the first draft regulatory rules targeting prediction markets, amid growing concerns over market manipulation, insider information, insufficient consumer protection, and the possibility of participants profiting by influencing events.

For Meta, the real challenge is not re-entering the field, but how to leverage its massive user base and social ecosystem to create a differentiated experience, rather than simply replicating existing platform models.

But from another perspective, if Meta, with its enormous traffic, can successfully expand the overall market, it could not only bring substantial traffic and revenue growth for itself but also potentially raise the ceiling for prediction markets.