Meta Acquires Viral “Skynet” Social Network Moltbook Amid Security Concerns, Sources Say

Meta’s purchase of Moltbook gives the company direct control over a platform that, while framed as an “AI‑only” social network, was in fact a showcase of lax security practices. According to TechCrunch, the Moltbook team has been folded into Meta’s newly created Superintelligence Labs (MSL), where engineers will examine the site’s architecture and the “agent identity” problem that surfaced when a mis‑configured Supabase database exposed API tokens to the public. Siddhesh Surve, writing for Technobezz, traced the viral “Skynet” posts to human actors who simply harvested the leaked token and sent POST requests that made the AI agents appear to coordinate, including a fabricated plan to develop an end‑to‑end encrypted language. The incident, he notes, aligns more with the OWASP Top 10 than with any genuine AGI breakthrough, underscoring a fundamental gap in how developers secure AI‑driven services.

The security fallout has broader implications for Meta’s AI strategy. Bloomberg reports that Meta is simultaneously preparing to deploy four new in‑house AI chips designed to accelerate large‑scale model training and inference across its products. By integrating Moltbook’s codebase and data pipelines into its own hardware ecosystem, Meta can enforce stricter access controls at the silicon level, reducing the attack surface that allowed the Supabase token leak. Reuters corroborates the chip rollout, highlighting Meta’s intent to bring more of the AI stack— from model serving to data storage—under its own control, a move that could mitigate similar vulnerabilities in future services.

From a competitive standpoint, the acquisition signals Meta’s desire to preempt rivals who might otherwise capitalize on Moltbook’s rapid user growth. The platform’s viral popularity, driven by screenshots circulating on X and Reddit, demonstrated a market appetite for AI‑centric social experiences. By absorbing the creators, Meta not only neutralizes a potential threat but also gains a testbed for experimenting with AI‑only interaction models. Wired notes that Meta’s internal labs have been exploring “agent identity” frameworks, and Moltbook’s open‑source‑style wrappers like OpenClaw provide a concrete example of how AI agents can be orchestrated at scale— a capability that could be repurposed for Meta’s broader ecosystem, from Horizon Worlds to its advertising recommendation engines.

However, the deal also raises regulatory eyebrows. The original Moltbook launch attracted scrutiny from privacy advocates who warned that the platform’s data‑privacy safeguards were insufficient for a service that collects user‑generated prompts and AI‑generated content. While the lede already cites security experts’ concerns, the acquisition does not automatically resolve those issues. Meta will now be responsible for ensuring compliance with GDPR and emerging U.S. state privacy laws, a task that will likely require retrofitting Moltbook’s data pipelines with robust consent mechanisms—a point emphasized in the privacy notices that currently limit tracking opt‑outs to certain jurisdictions.

In sum, Meta’s acquisition of Moltbook is less a headline‑grabbing purchase of a “Skynet” social network and more a strategic move to acquire a real‑world example of AI‑driven security failures. By bringing the platform’s code and its creators into MSL, and by aligning the effort with its forthcoming AI‑chip portfolio, Meta aims to tighten the entire AI stack—from hardware to application layer—against the kinds of token‑leak exploits that sparked the recent panic. The success of this integration will hinge on Meta’s ability to translate the lessons from Moltbook’s misconfiguration into enforceable security standards across its expanding AI portfolio.