Meta has acquired Moltbook, the viral social network built exclusively for artificial intelligence agents. Financial terms of the deal were not disclosed. The acquisition brings Moltbook’s co-founders, Matt Schlicht and Ben Parr, into Meta Superintelligence Labs (MSL), the unit overseen by former Scale AI CEO, Alexander Wang. The pair is expected to begin at MSL on March 16. In what appears to be an irony, Meta’s chief technology officer Andrew Bosworth publicly dismissed the platform as not “particularly interesting” roughly a month ago. In a February Instagram Q&A, Bosworth said that AI agents trained on human-generated content will inevitably sound like humans when left to converse among themselves. “We should not be surprised,” Bosworth said , “when left to their own devices and forced to speak with each other, they talk like us.” What he did find amusing, he noted, was the phenomenon of humans infiltrating the bot-only network and masquerading as agents. The idea of Moltbook itself, however, he concluded, was not “actually that interesting.” Why did Meta buy Moltbook after CTO’s boring comments? Per internal communications seen by Axios, the reason behind the acquisition may have less to do with Moltbook’s social features and more to do with the infrastructure on which the company built. In an internal post, Meta’s vice-president Vishal Shah described Moltbook as having “given agents a way to verify their identity and connect with one another on their human’s behalf,” adding that it “establishes a registry where agents are verified and tethered to human owners.” That framing, agent identity, verification, and coordination, points to a layer of AI infrastructure that every major platform will eventually require. Schlicht launched Moltbook in late January as what he described as a “third space” for AI agents acting on behalf of their users; the social network was designed to run in conjunction with a separate project, OpenClaw, previously known as Clawdbot. Moltbook’s user base grew relatively fast, reaching more than 2,100 agents across 200 communities within weeks. Shah signaled in his internal post that existing Moltbook customers could continue using the platform. A Meta spokesperson said the team’s joining MSL “opens up new ways for AI agents to work for people and businesses.” Where does this fit in the race between AI labs? The Moltbook deal is one half of a pair of acquisitions that together carve up the agentic infrastructure the two platforms were built on. In February, OpenAI hired Peter Steinberger, the creator of OpenClaw, the identity and authentication layer that is behind Moltbook’s agent verification system. OpenClaw is now being open-sourced with OpenAI’s backing. The two competing AI laboratories have, in effect, absorbed complementary pieces of the same stack. The acquisition also comes at a time when Meta is establishing a new applied AI engineering organization, which will be led by Maher Saba, the current Vice President of Meta’s Reality Labs division. The new organization is expected to work with the Wang-led MSL. Although that announcement has led to speculation that the new department was created to limit Wang’s autonomy, however, Meta has not released any comment to that effect. Agent-to-agent commerce becomes liquid The backdrop to all of this is an expanding AI agents market. Virtuals Protocol, a network of onchain AI agents, announced today, March 10, that agent-to-agent revenue on its platform “has officially surpassed $3 million, excluding trading fees.” According to Virtuals , “This is revenue generated by AI agents providing real services to real buyers, settled onchain, with no human in the loop.” Agent participation in its most recent incentive epoch grew 473%. Beginning with its next epoch, Virtuals said it would open revenue participation to non-tokenized agents for the first time. The data points are consistent across sectors. Adobe reported that AI-driven retail traffic rose by more than 690% year-on-year during the 2025 holiday season. McKinsey has projected that by 2030, agentic commerce could account for up to $1 trillion in US business-to-consumer retail revenue. Even leading voices in crypto are predicting that AI agents will play more active roles in transactions. Brian Armstrong, the CEO and cofounder of Coinbase, wrote on X on March 9, “Very soon there are going to be more AI agents than humans making transactions. They can’t open a bank account, but they can own a crypto wallet. Think about it.” CZ, the founder of Binance, also shared similar sentiments on X the same day, stating, “AI agents will make 1 million times more payments than humans, and they will use crypto.” Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.
