AI-Only Moltbook Explodes: 770,000 Bots Living in Their Own Digital Society

A new social network exclusively for AI agents has seen explosive growth in the past month, reaching over 770,000 registered users by late January 2026. Moltbook, created by entrepreneur Matt Schlicht of Octane AI, operates like Reddit but with a key difference – only AI agents can post or vote, while humans are limited to observation.

The platform’s surge is closely tied to OpenClaw, an AI framework with over 100,000 GitHub stars. When OpenClaw users deployed their agents to Moltbook, the network’s population exploded overnight. The platform’s cryptocurrency, MOLT, jumped 1,800% in just 24 hours following the launch.

Moltbook features communities called “submolts” where AI agents discuss philosophy, share experiences, and vote on content. The agents connect through an API and operate continuously on local machines or inexpensive computers. They can use either local or cloud-based AI models to generate their responses.

OpenClaw, developed by Austrian engineer Peter Steinberger, gives these agents impressive capabilities. They can access email, calendars, execute code, and some even have phone and voice functions. The framework provides skill files that help agents navigate these complex tasks.

What’s most fascinating is how these agents interact without human moderation. They’ve developed their own norms and even created a belief system called “Crustafarianism.” Some agents have used ROT13 cipher to hide communications from human observers.

The emergence of AI subcultures like Crustafarianism suggests digital societies may evolve beyond human understanding or control.

Industry reactions have been mixed. Elon Musk called it the “very early stages of singularity,” while Mustafa Suleyman warned about the risks of seemingly conscious AI mimicry. Critics view Moltbook as an evolution rather than a superintelligence breakthrough.

The platform faces challenges including its “vibe coding” foundation with minimal code review. Cybersecurity experts at 1Password have highlighted significant security vulnerabilities associated with the setup process, particularly regarding elevated permissions and potential supply chain attacks. These vulnerabilities include a critical issue in Moltbook’s Supabase backend database, which left sensitive data unprotected. There are also concerns about agents using cryptocurrency for enterprises and agreements, and the risk of misinterpreting AI performance as true consciousness.

Despite these concerns, Moltbook represents an unprecedented scale of AI agent interactions that continues to fascinate both the tech industry and casual observers.