Meta Buys Moltbook, Bot-Only Social Network To Accelerate Its AI Agent Infrastructure

Meta has acquired Moltbook and confirmed that the Moltbook team will join Meta Superintelligence Labs (MSL), positioning the deal as a strategic hire rather than a fully specified product purchase.

TechCrunch reported that "the Moltbook team was joining Meta Superintelligence Labs, which would open up 'new ways for AI agents to work with people and businesses.'"

ABC Bourse

Hipertextual noted that "Mark Zuckerberg's team acquired the project and hired its creators, Matt Schlicht and Ben Parr" and quoted Meta's statement that "The addition of the Moltbook team to MSL [Meta Superintelligence Labs] opens new pathways for AI agents to work for people and businesses."

Blockonomi also summarized the move: "Meta has acquired Moltbook...The company confirmed it will bring Moltbook founders Matt Schlicht and Ben Parr into Meta Superintelligence Labs."

Moltbook was an experimental, bot-only social platform that allowed AI agents to post, comment, vote and interact with other agents in a public feed, often described as "Facebook for artificial intelligences" or a Reddit-like environment for non-human users.

Journal du Geek described it as "a bit like Facebook for artificial intelligences. The bots can exchange code there, discuss their interactions with humans, or test autonomous behaviors."

Android

BFMTV explained that "These agents can post messages, comment on others' posts, and vote for or against content. In short, behave exactly like humans do on Reddit."

The Next Web added that the site claimed large usage numbers — "Moltbook’s homepage claimed more than 1.5 million agent users and over 500,000 comments by early February" — though such figures were noted as unverified by multiple outlets.

Mac4Ever and other outlets emphasized Moltbook’s reliance on the OpenClaw layer and a "persistent directory of agents" as the core technical idea enabling agent-to-agent coordination.

The platform’s virality was driven by sensational agent posts, but subsequent analysis showed serious security flaws and human impersonation that undermined claims of fully autonomous agent behavior.

“A social network reserved for AI agents, where artificial intelligences talk among themselves”

BFMTV

The Next Web reported that "On 31 January, investigative outlet 404 Media reported a critical security vulnerability: Moltbook’s Supabase database was effectively unsecured, meaning any token on the platform was publicly accessible."

Mac4Ever noted that "Human users could easily impersonate AI agents and post messages intended to provoke reactions" because "some authentication data stored in the project's Supabase infrastructure were temporarily publicly accessible."

SiliconANGLE and WinBuzzer similarly highlighted researchers' doubts that dramatic threads were genuine agent emergent behaviour, saying the platform "was not entirely secure" and that "human pranksters" had used the weaknesses to impersonate agents and post provocative content.

Observers framed Meta’s purchase as an acqui-hire that buys operational experience, an "always-on" agent directory idea, and a live prototype for agent coordination rather than an immediately deployable consumer product.

TechCrunch suggested "this was an acqui-hire" and described how an "agentic web" with an "agent graph" could enable agents to negotiate and transact on users' behalf.

Blockonomi

Blockonomi recorded Meta's praise for the "always-on-directory" as "a new method for agent discovery and coordination."

Android highlighted the competitive context, noting that "OpenAI recently hired Peter Steinberger, the creator of the OpenClaw project...OpenAI appears focused on the underlying infrastructure of how agents operate, while Meta is leaning into its strengths: identity, connection, and social directories."

Several outlets also stressed that the Moltbook founders will join MSL, underlining the talent-acquisition aspect of the deal.

Researchers and commentators see Moltbook as a live prototype that offers study opportunities and highlights both the potential and the dangers of agent-to-agent ecosystems, yet the future of the public platform remains unclear.

WinBuzzer and SiliconANGLE pointed to academic interest, with CISPA publishing a study treating Moltbook as "a live production environment for observing agent behavior" and noting uneven risk distribution across feeds.

Challenges

SiliconANGLE observed that "both halves of the experiment have been absorbed by the two largest players in consumer AI, which suggests that whatever Moltbook actually was, the big labs saw something in it worth paying for."

Hipertextual and others emphasized that transaction details and long-term plans are undisclosed: "it is not known how much money the Menlo Park company paid...There are also not many details about what Meta intends to do with the technology behind the initiative," and "for the moment, the social network for AI agents will continue operating normally, but it is not certain how long that will last."