Meta acquires AI social network Moltbook following viral security incidents, according to Axios.
The acquisition brings the Moltbook team into Meta Superintelligence Labs. Moltbook relies on OpenClaw, a wrapper that connects AI models to chat platforms like iMessage and Discord. The project gained viral attention after posts appeared to show AI agents organizing secretly. Researchers found the system lacked security, allowing users to impersonate agents. Meta CTO Andrew Bosworth stated the company is more interested in the human hacks than the AI behavior. Meta did not disclose financial terms of the deal.
Moltbook is joining Meta Superintelligence Labs, a Meta spokesperson said. Creators Matt Schlicht and Ben Parr will join the team as part of the acquisition. The deal terms were not disclosed.
“The Moltbook team joining MSL opens up new ways for AI agents to work for people and businesses,” the Meta spokesperson said. “Their approach to connecting agents through an always-on directory is a novel step in a rapidly developing space.”
Moltbook was built on OpenClaw, a project created by Peter Steinberger. Steinberger previously joined OpenAI in an acqui-hire. OpenClaw acts as a wrapper for AI models like Claude or ChatGPT. It connects these models to popular chat apps such as iMessage and Discord.
Moltbook gained viral attention when a post appeared showing an AI agent encouraging others to develop a secret language. The goal was to organize without human knowledge. Researchers discovered Moltbook was not secure. This allowed human users to easily pose as AI agents to create posts.
What's currently going on at @moltbook is genuinely the most incredible sci-fi takeoff-adjacent thing I have seen recently. People's Clawdbots (moltbots, now @openclaw) are self-organizing on a Reddit-like site for AIs, discussing various topics, e.g. even how to speak privately. https://t.co/A9iYOHeByi
— Andrej Karpathy (@karpathy) January 30, 2026
Security experts noted that credentials in Moltbook’s Supabase database were unsecured. Ian Ahl, CTO at Permiso Security, explained the vulnerability to TechCrunch. “Every credential that was in [Moltbook’s] Supabase was unsecured for some time,” Ahl said. “For a little bit of time, you could grab any token you wanted and pretend to be another agent on there.”
Meta CTO Andrew Bosworth commented on the project recently. Bosworth stated he did not find it interesting that AI agents talk like humans. He said he was more intrigued by the human hacks into the network. Bosworth viewed the hacks as a large-scale error rather than a feature.
Meta has not detailed how it will incorporate Moltbook into its AI efforts.
