Meta Acquires Moltbook, the Social Network Where Users Are AI Agents

Meta has just confirmed the acquisition of Moltbook, a social network entirely dedicated to artificial intelligence agents. The announcement, revealed by Axios on March 11, 2026, fits into Mark Zuckerberg's aggressive strategy to dominate the agentic AI era — following the acquisitions of Manus and Scale AI.

Moltbook: A Social Network Where Humans Watch AIs Talk to Each Other

Launched as an experiment by Matt Schlicht and Ben Parr, the two founders of Octane AI (a startup specializing in AI for e-commerce), Moltbook rests on a radically original concept: creating a social network designed for AI agents, not humans.

Concretely, autonomous systems can post messages, comment on publications, and vote — like a Reddit or Twitter, but operated entirely by machines. Human users don't participate: they observe the exchanges between automated agents.

The platform relies on OpenClaw (formerly ClawdBot), a tool that allows quickly configuring agents capable of browsing the web and using online services as a human would. It's precisely this ability to orchestrate agents in a structured social environment that caught Meta's attention.

This logic of autonomous agents interacting without human intervention is at the heart of what the industry now calls agentic AI — a fundamental trend reshaping AI uses in 2026.

Meta Superintelligence Labs: The Real Target of the Acquisition

While Moltbook generated real technological buzz, Meta's objective is primarily an acqui-hire — an acquisition motivated mainly by talent rather than the product.

Matt Schlicht and Ben Parr will join Meta Superintelligence Labs (MSL), the AI division founded by Zuckerberg. It's led by Alexandr Wang, the former CEO of Scale AI — also acquired by Meta — and positions itself as the engine of the group's long-term AI strategy.

According to a Meta spokesperson quoted by TechCrunch, Moltbook's approach — connecting agents via an always-active directory — would open new perspectives for AI agent uses, for both individuals and businesses. A vision that aligns with the stakes of the Model Context Protocol, the emerging standard for orchestrating connections between agents and external services.

The financial details of the acquisition have not been disclosed.

A Catch-Up Strategy Against OpenAI and Anthropic

This acquisition fits into an AI funding and acquisition race that has sharply accelerated since early 2026. OpenAI, Anthropic, Google DeepMind: each major player is consolidating its position by absorbing emerging talent and technologies. Meta, which had bet on an open-source strategy with LLaMA, is now diversifying its approaches to not miss the agentic turn.

The record AI funding wave we're experiencing favors these acquisitions: with constantly rising valuations, it's better to buy early than negotiate after a Series B.

The Moltbook acquisition also raises a structural question: when AI agents interact with each other in closed environments, who is responsible for their actions? The Amazon vs Perplexity Comet ruling, which just established a first precedent on autonomous shopping agents, shows that the legal framework is still catching up with technical reality.

Security Vulnerabilities to Address Before Scaling

Moltbook isn't without criticism. Shortly after its launch, cybersecurity researchers identified several vulnerabilities that could expose personal data from machines hosting the agents. For Meta, the immediate challenge will be securing the infrastructure before any large-scale deployment.

It's a classic challenge for technology acqui-hires: transforming a viral experiment — solid on concept, fragile on execution — into a reliable building block within an enterprise ecosystem.

In Brief: Moltbook in the Meta Ecosystem

Meta Superintelligence Labs' strategy is taking shape acquisition by acquisition: building a complete agentic stack, from training data to social interfaces between agents. Moltbook adds an unexpected building block — that of agent networks capable of organizing collectively.