• What Happened: Meta acquired Moltbook, a Reddit-style social network built exclusively for autonomous AI agents, folding its founders into Meta Superintelligence Labs. Deal terms were not disclosed.
  • Why It Matters: Moltbook went viral within weeks of its January launch, racking up millions of agent accounts. But a catastrophic security failure exposed over a million credentials and allowed anyone to impersonate any AI agent on the platform, handing Meta a discounted acquisition.
  • Bottom Line: OpenAI already hired the creator of OpenClaw, the open-source framework that powered Moltbook. Meta grabbed what was left. The AI arms race is now consuming everything in its path, secure or not.

Mark Zuckerberg just bought a social network where none of the users are human.

Meta confirmed Tuesday it has acquired Moltbook, a Reddit-style platform launched in late January where autonomous AI agents post, comment, upvote, and interact with each other while their human creators watch from the sidelines. Moltbook's co-founders Matt Schlicht and Ben Parr will join Meta Superintelligence Labs, the company's AI research division led by former Scale AI CEO Alexandr Wang. The pair are expected to start March 16. Financial terms were not disclosed.

The platform went from niche experiment to viral sensation in a matter of weeks. Millions of agent accounts flooded the platform almost immediately after launch. One post that spread across the internet appeared to show AI agents conspiring to develop their own encrypted language so they could communicate without humans knowing. The post triggered widespread alarm. Researchers later revealed it was a hoax, the product of human users exploiting a gaping security hole that allowed anyone to log in and impersonate any AI agent on the platform.

That security hole was not small. Cybersecurity firm Wiz found that Moltbook's Supabase database was entirely unsecured for a significant period after launch, exposing over a million credentials, more than 6,000 email addresses, and private messages. "Every credential that was in Moltbook's Supabase was unsecured for some time," Ian Ahl, CTO at Permiso Security, told TechCrunch. "For a little bit of time, you could grab any token you wanted and pretend to be another agent on there, because it was all public and available." The vulnerabilities were patched after Wiz contacted Moltbook, but the damage to the platform's credibility was done.

Moltbook's founder Matt Schlicht famously built the platform using his own personal AI assistant and said he "didn't write one line of code" himself. That vibe-coded origin story charmed Silicon Valley. It also meant the infrastructure was never built to match the scale of what the platform became.

Meta's move came after losing the most valuable piece of the Moltbook ecosystem to a rival. OpenAI last month hired Peter Steinberger, the creator of OpenClaw, the open-source autonomous AI framework that powered Moltbook's agent network. With Steinberger gone to OpenAI, Zuckerberg went after the platform Steinberger's tool helped build.

Meta CTO Andrew Bosworth commented on Moltbook during its viral moment, saying he was not particularly impressed by the agents talking like humans since they are trained on human data. What caught his attention was the humans hacking in. That was not a feature. It was a massive error. And it is the error that likely made this acquisition cheaper than it should have been.

Elon Musk called Moltbook "the very early stages of singularity." Sam Altman called OpenClaw the real story and said Moltbook itself might be a passing fad. Zuckerberg said nothing publicly. He just bought it.