Zuckerberg directs Meta to build a prediction markets app called Arena to rival Polymarket and Kalshi

Meta's Prediction Markets Play: A Defensive Gamble on Sentiment

Mark Zuckerberg has directed Meta to build a prediction markets app internally called Arena, according to reporting from Mike Isaac. The experimental app would operate independently of Facebook and Instagram and compete directly with Polymarket and Kalshi, the dominant platforms in the space.

The app currently will not use real money, though insiders have not fully ruled out monetization in the future. This constraint matters. Without financial stakes, Meta sidesteps the regulatory and reputational risks that have dogged prediction markets—the addiction lawsuits, the fraud scandals, the sense that these platforms prey on the vulnerable. Kalshi alone reportedly generates over $2 billion in annualized revenue and is growing fast, but sentiment around the category has deteriorated significantly. Walter Pohl's recent exposé on Polymarket's advertising practices and the shift from political betting toward sports gambling have left the category's cultural standing damaged.

The strategic logic behind Arena is less clear. Sentiment on Meta itself is near rock bottom among both the capital markets and its own employee base. Zuckerberg may simply be operating under the assumption that things cannot get worse—a posture of "f it, let's ship it." The company is simultaneously redirecting 30 to 50% of its core engineering teams toward data labeling for AI training, a move that has reportedly left employees feeling randomly conscripted, like tributes in The Hunger Games.

One potential steel man: if Arena can make the status game of making accurate predictions as rewarding as viral Instagram posts—if you can browse someone's profile and see their follower count, their reach, and their Arena track record all at once—then Meta might build a product people genuinely use, even without money changing hands. The company has never monetized engagement through pay-to-play mechanics; it has gamified everything else in its orbit (likes, followers, story views) without betting real capital. That model could work here.

But the counterargument is starker. Kylie Jenner's Meta glasses partnership, likely valued around $50 million, has driven genuine enthusiasm. A $100 million Robert Downey Jr. deal with Snap for the same purpose reads as irrational by comparison. Kylie drives purchases; RDJ drives brand awareness at best. More fundamentally, Meta is betting hundreds of billions on AI dominance while introducing gambling mechanics to the most addictive social platform on earth—a move that invites the exact criticism already leveled against the platform in ongoing addiction litigation.

If Arena stays free forever, it's a different animal than sports betting. If the money switch flips, the PR catastrophe is baked in. The market has not priced this as a material business opportunity; it reads more as a distraction or a bet on internal culture than as a revenue driver.