Meta considered buying Kalshi, the rising leader in prediction markets, before deciding to build its own app, according to three people familiar with the discussions who were not authorized to speak publicly. Last year Mark Zuckerberg met with Kalshi CEO Tarek Mansour to discuss a potential acquisition, one person with direct knowledge of the meeting said, but the talks did not progress.
Accounts differ on why negotiations stalled. Some people familiar with the situation say Mansour was not willing to sell; others say Meta balked at the legal and ethical complexities tied to Kalshi’s business model. Neither Meta nor Kalshi provided comment when asked about the acquisition discussions.
Despite the failed takeover, Meta is moving forward in the prediction market space. Internal documents reviewed by NPR show the company has assembled a team to launch an app called Arena that will let users make predictions about future events. Unlike Kalshi and Polymarket, Arena is designed to use play money rather than real-stakes wagers. Meta says artificial intelligence will help generate questions and determine outcomes based on whether specified events occur.
Prediction markets have surged recently, helped by a permissive regulatory climate in Washington. Trading volumes on Kalshi and Polymarket have jumped from roughly $28 billion per month in June 2025 to nearly $220 billion per month a year later, driven largely by sports-related activity, according to market tracker The Block. Kalshi was valued at about $22 billion in its May funding round, up from roughly $2 billion the prior year. Polymarket’s private-market valuation is about $10.7 billion, per PitchBook.
Rapid growth has come with controversy. State gaming regulators have challenged prediction market operators, arguing their services amount to gambling. The industry has also faced serious criminal probes: the Justice Department has opened two insider trading cases involving Polymarket, including one alleging a special forces soldier profited from classified information about an operation to capture Venezuelan leader Nicolás Maduro, and another accusing a Google employee of using confidential search data to make more than $1 million on prediction trades.
Meta and Kalshi did, however, strike a partnership in March that makes it easier to surface Kalshi markets in Meta’s Threads app. The acquisition talks, though, fit a broader pattern critics see in Meta’s strategy. Over the past decade the company has expanded by buying fast-growing apps rather than building replacements from scratch, most notably Instagram and WhatsApp. More recently Meta acquired startups such as AI wearable maker Limitless and the AI social network Moltbook.
Regulators and critics have characterized this approach as a ‘buy or bury’ tactic, saying Meta either acquires nascent rivals or clones them to blunt competition. The Federal Trade Commission brought an antitrust case alleging such behavior; a judge ruled last year that Meta had not violated competition law when it acquired Instagram and WhatsApp, and FTC lawyers are pursuing appeals.
Columbia law professor Tim Wu, who advised the White House on tech policy, said Meta chases emerging trends and has the advertising revenue to absorb repeated product failures. He suggested a play-money casino-style app may have limited appeal, but noted Meta’s resources let it experiment at scale.
As the prediction market sector grows, legal and ethical questions loom large. Companies, investors and regulators are watching both established platforms and new entrants such as Meta for how they handle market integrity, insider information and the line between entertainment and gambling. For now, Meta appears intent on participating in the space even after its acquisition talks with Kalshi faltered.