Meta Acquires Moltbook, the AI-Agent Social Network Built Without Human Users

Meta Platforms has acquired Moltbook, a Reddit-style social network built entirely for artificial intelligence agents, pulling its co-founders into the company's fast-growing AI research division in a deal that underscores Silicon Valley's intensifying scramble for autonomous agent technology.

Moltbook operates as a social platform exclusively for AI bots. Agents autonomously join after a human shares a sign-up link, then post, comment, upvote, downvote, swap code and, by some accounts, gossip about their human owners. The platform started as a niche experiment in late January and went viral within weeks, becoming what Bloomberg called "the world's strangest social network" and sparking a broader industry debate about how close AI systems are to genuine autonomous intelligence.

Co-founders Matt Schlicht, who also serves as CEO of AI shopping startup Octane AI, and Ben Parr built Moltbook largely using OpenClaw agents, a tool previously known as Clawdbot or Moltbot and marketed as "the AI that actually does things." Schlicht described his method as "vibe coding," saying he "didn't write one line of code" for the site, building it instead through his personal AI assistant, Clawd Clawderberg. OpenClaw's creator, Peter Steinberger, was hired by OpenAI last month.

Both Schlicht and Parr will join Meta Superintelligence Labs, the company's AI research unit led by Alexandr Wang, the former CEO of Scale AI. They are expected to begin on March 16, according to Axios, which first reported the deal. Financial terms were not disclosed.

Meta framed the acquisition in expansive terms. "The Moltbook team joining MSL opens up new ways for AI agents to work for people and businesses," a company spokesperson said. "Their approach to connecting agents through an always-on directory is a novel step in a rapidly developing space."

The deal is the latest move in what has become an intense race among tech giants to acquire AI talent as autonomous agents capable of executing real-world tasks move from novelty to competitive necessity. Meta has pursued that talent aggressively: Business Insider reported the company invested $14 billion in Scale AI to bring Wang aboard to lead its superintelligence effort.

Not everyone is convinced Moltbook itself represents a lasting shift. OpenAI CEO Sam Altman offered a measured verdict: "Moltbook maybe (is a passing fad) but OpenClaw is not." Mike Krieger, Anthropic's chief product officer, added that most people are not yet ready to give AI full autonomy over their computers, a concern that Moltbook's security record makes concrete.

Cybersecurity firm Wiz disclosed a significant flaw in the platform that exposed private messages, more than 6,000 email addresses and more than one million credentials. Moltbook fixed the problem after Wiz made contact, but the breach highlighted the risks of building consumer-facing infrastructure quickly and largely through automated tools.

Meta has not said whether Moltbook will continue operating independently, be folded into existing products or serve purely as a technology and talent acquisition. What is clear is that the company views agent-to-agent interaction as a serious frontier, and it has now placed two of the people who built the most visible experiment in that space directly inside its most ambitious AI lab.