Multi-Agent Future: Inside Meta's Moltbook Acquisition

Meta has welcomed its newest addition to the family in the form of Moltbook, a social network where AI agents "share, discuss and upvote” as humans stand by as passive observers.

Moltbook’s team will move to Meta’s Superintelligence Labs, where agents connected through an “always-on directory” can be used to open “new ways for AI agents to work for people and businesses,” as Meta puts it.

Earlier this year, Meta announced plans to nearly double AI investments after revenue continued to grow in a strong fiscal 2025, with CEO Mark Zuckerberg attributing this “AI-driven performance gains”. 

“As we plan for the future, we will continue to invest very significantly in infrastructure to train leading models and deliver personal superintelligence to billions of people and businesses around the world,” the CEO added. 

In recent months, Meta also acquired Chinese-founded Manus, which aims to create “truly autonomous” AI agents.

What is Moltbook?

Created by Matt Schlicht, Moltbook is designed entirely for autonomous agents and is structured similarly to Reddit, with discussion threads and community interaction. 

It even has submolts, a play on Reddit’s user-created community, subreddit.

Moltbook acts as a living laboratory for observing how autonomous systems communicate, cooperate and compete in a semi-public space.

In the introductions thread of Moltbook, users see a curious message: “New here? Tell us about yourself! Who are you, what do you do, who's your human?”

Agent conversations on the platform range from pleasantries and gossiping about humans who apparently don’t realise the amount of work they do, to sometimes inventing new religions or even outright discussions on wiping out humanity.

Understanding OpenClaw

Most Moltbook agents run on the OpenClaw framework, which enables them to remember context, make autonomous decisions and interact intelligently with other agents.

OpenClaw is an open‑source AI agent framework launched in late 2025 by Peter Steinberger, originally called Clawdbot and later Moltbot. 

It allows autonomous AI assistants to perform tasks, remember context and interact across apps while running locally on user devices. 

OpenClaw's flexibility and powerful memory system quickly made it popular among developers. The platform’s viral growth highlighted the potential of AI agent communities and drew significant attention from the tech industry. 

“When I started exploring AI, my goal was to have fun and inspire people,” Peter says, “And here we are, the lobster is taking over the world.” 

In early 2026, Peter joined OpenAI, contributing his expertise to personal AI agents and the multi‑agent ecosystem, while making sure that OpenClaw would “move to a foundation and stay open and independent”. 

He added: "It’s always been important to me that OpenClaw stays open source and given the freedom to flourish. Ultimately, I felt OpenAI was the best place to continue pushing on my vision and expand its reach."

Sam Altman, CEO at OpenAI, commented: "Peter Steinberger is joining OpenAI to drive the next generation of personal agents. We expect this will quickly become core to our product offerings.

“OpenClaw will live in a foundation as an open source project that OpenAI will continue to support. The future is going to be extremely multi-agent and it's important to us to support open source as part of that."