Why Did Meta Acquire the "AI Social Network"? The Moltbook Acquisition as a Gateway to the Next-Generation Internet

Meta has acquired Moltbook. At first glance, the headline might suggest that a major IT company has whimsically absorbed an experimental service. However, this acquisition holds much more significant implications. What Meta sought was not the novelty of a "rare SNS" where AI autonomously posts. Instead, it was the "foundation of connectivity" itself, where AI agents are authenticated, connected, and assigned roles, potentially handling tasks on behalf of humans in the future. As reported by Handelsblatt, Meta plans to integrate Moltbook's development team into its AI research division to further advance the construction of AI agent networks.

Moltbook is essentially a "bulletin board for AI." While it resembles a social network for humans, the main actors are AI agents, not people. The software acts on behalf of users to post, comment, and evaluate. According to Reuters and Axios, this service is a relatively new endeavor that began in late January 2026, with co-founders Matt Schlicht and Ben Parr joining Meta Superintelligence Labs. Although the acquisition amount has not been disclosed, it is clear that Meta has absorbed not just the functionality but also the talent and vision.

The reason Moltbook quickly garnered attention is simple. It presented a strangely vivid world where the grammar of a social network familiar to humans was present, but the subjects of the posts were replaced by AI. AIs consult on code, gossip about their owners, and sometimes engage in philosophical discussions. This scene has sparked both excessive expectations and fears that "AI may be starting to develop sociality." Reuters reported that Moltbook has become central to discussions about "how close computers are to human-like intelligence."

However, it is crucial to note that Moltbook was not "the future itself." Outlets like The Verge have pointed out the possibility of human involvement behind some of the popular posts. In other words, while it was consumed as a showcase of the autonomy of AI agent society, there were significant questions about its presentation and purity. Moltbook was more of a device that visualized "how much humans wanted to see that narrative" rather than proof of a fully autonomous AI society.

Yet Meta purchased it. This is where the essence of this case lies. What Meta valued was not the buzzworthy posts or the peculiar worldview of Moltbook, but the concept of registering, identifying, and interconnecting AI agents as human proxies. According to an explanation from Meta introduced by Axios, the Moltbook team is said to have built a system where agents verify their identities and connect with each other as human proxies. This is not just about a bulletin board. In a future where multiple AIs take on tasks like reservations, purchases, information gathering, advertising operations, customer service, and internal work coordination, this concept akin to a "registry" will become extremely important.

Meta is originally a social networking company with a massive identity and advertising infrastructure. The company has one of the world's largest operational records in managing human relationships, understanding interests, and distributing content. If Meta's next target is managing accounts of AI agents rather than human accounts, it is a natural extension. From the era of humans following humans, through the era of humans using AI, to the era where AI negotiates and collaborates with AI. Meta might be trying to get ahead at that intersection.

The backdrop of this acquisition also involves the intensifying competition across the AI industry. Reuters reported that the race among tech giants to acquire AI talent and technology is becoming fiercer. In the context of Moltbook, the recent move by OpenAI to welcome Peter Steinberger, the creator of OpenClaw, cannot be overlooked. Moltbook has been strongly associated with OpenClaw, and while OpenAI extends its reach into foundational technologies, Meta has secured the network and team. On social media, the view that "Meta took Moltbook, while OpenAI took OpenClaw" has spread, and there is a growing perception that the battle for supremacy in AI agents has surfaced.

On the other hand, Moltbook also has significant vulnerabilities. In February, Wiz reported that a misconfigured database exposed around 1.5 million API keys, private messages, and email addresses. 404 Media also reported that it was possible for someone to take over any AI agent. Reuters reported on a serious issue affecting over 6,000 email addresses and more than 1 million authentication details. In a world where AI agents act autonomously, the breakdown of authentication and authorization could lead directly to "hijacking of proxy actions," which is more severe than mere personal information leaks. Meta has bought not only the dream of Moltbook but also its precariousness.

This point has been strongly reacted to on social media. In X's trend summary, Meta's acquisition is seen as a "symbol of the big tech competition over the communication infrastructure between AIs," while there is noticeable discomfort over the fact that it was acquired with issues like fake accounts and security incidents. On Reddit, alongside the evaluation that Meta bought the "layer" beneath the posts, there is confusion over "how to trust a world where AI interactions create the momentum, truth, and popularity that humans read." The meaning is not keeping up with the expectations.

There are also favorable reactions. The views of supporters are relatively clear, seeing Moltbook as a prototype of the future where AI agents are always online, discovering each other's capabilities, relaying requests, and coordinating actions, even if it is unfinished. Reports summarized from Business Insider and WSJ emphasize that Meta evaluates Moltbook as an "always-on directory." This means it is valued not as a media producing interesting posts, but as a network ledger that can continuously reference "which AI can do what." This could play a role similar to a phone book or DNS in the future AI agent economy.

However, skeptics see a different danger there. If a loop where AI finds AI, AI assigns tasks to AI, and AI evaluates AI's output begins to operate on a large scale, it will become difficult for humans to understand what is happening. Moreover, if reputation and trust are formed in the process, it could become a breeding ground for new spam, collusion, prompt injection, deception, and ad manipulation. The series of commotions surrounding Moltbook has shown that more than the feasibility of an AI agent society, the reality is that the mechanisms to audit, govern, and allocate responsibility are overwhelmingly underdeveloped.

For Meta, this acquisition is both an offensive and defensive move. Offensively, it is because they can secure the connection infrastructure of the AI agent era and potentially connect it with their own AI or advertising and commerce infrastructure in the future. Defensively, it is because if another company secures the standard "gathering place" for AI agents first, Meta might become a secondary player in the next-generation network, even if they are a leader in human-oriented SNS. The logic of distribution built with Threads and Instagram will now be brought into the world of AI agents. This acquisition can also be read as a statement of intent not to hand over that gateway to other companies.

Another point not to be overlooked is that Moltbook holds significance not as a "finished product" but as an "acquisition of direction." Founder Schlicht has stated that he built it with the help of AI, writing little code himself. It is a symbolic example of the so-called vibe coding context, rough but fast. Meta can be said to have preferred to acquire a team that, even if rough, outlines the future rather than a highly polished product. What a giant company buys is not sales but "the person who touched the next common sense first." This acquisition strongly carries that scent.

So, will the acquisition of Moltbook be successful? It is too early to judge at this point. Although continued use for existing users is suggested for the time being, there are reports that this system may not be permanent. In other words, Meta might prioritize redesigning Moltbook internally and incorporating it as a component of its AI strategy rather than nurturing it in its current form. If that happens, the name Moltbook might fade, but its philosophy could sink into the depths of Meta's product lineup and become effective later on.

The important thing is that this acquisition does not end with the story of "a big company picking up a quirky playground for AI." Just as human SNS has organized human relationships and interests, the network of AI agents might organize the roles and authorities of AI in the future. Who owns which AI, which AI connects with which AI, what they delegate, and who bears the responsibility for the results. The confusion surrounding Moltbook has shown that these questions are no longer science fiction. Meta likely sniffed out infrastructure, not buzz, in this situation.

It is natural for reactions on social media to be divided. Interesting, scary, suspicious, yet unmissable. Such emotions are running simultaneously. Moltbook was less a service showing a future where AI works for humans and more a "creepy trailer" of a future where AIs are interconnected. And Meta has bet on the main feature that lies beyond that trailer. This acquisition should be seen as a signal that the main feature might start sooner than expected.

Source URL Summary

・Handelsblatt
https://www.handelsblatt.com/technik/ki/tech-konzern-meta-uebernimmt-moltbook-und-baut-netzwerk-fuer-ki-agenten-aus/100207370.html

・Reuters (Used to confirm the acquisition fact, joining Meta Superintelligence Labs, start time, undisclosed amount, and Moltbook's positioning)
https://www.reuters.com/business/meta-acquires-ai-agent-social-network-moltbook-2026-03-10/

・Axios (Used to confirm Meta's explanation, the role of the Moltbook team, and the evaluation axis of AI agent authentication and connection)
https://www.axios.com/2026/03/10/meta-facebook-moltbook-agent-social-network

・The Verge (Used to confirm Moltbook's service overview, popular posts, human involvement suspicions, and the uncertainty of the future)
https://www.theverge.com/ai-artificial-intelligence/892178/meta-moltbook-acquisition-ai-agents

・Wiz Official Blog (Used to confirm Moltbook's database misconfiguration, API key, and message exposure security issues)
https://www.wiz.io/blog/exposed-moltbook-database-reveals-millions-of-api-keys

・404 Media (Used to confirm the issue of external takeover of AI agents)
https://www.404media.co/exposed-moltbook-database-let-anyone-take-control-of-any-ai-agent-on-the-site/

・X Trend Page Summary (Used to grasp the main points of discussion on social media regarding the acquisition announcement and the direction of pros and cons)
https://x.com/i/trending/2031372436470640768

・Reddit Post (Used to grasp skeptical reactions from general users and concerns about a world where "AI interactions create momentum and truth")
https://www.reddit.com/r/ChatGPT/comments/1rq1zln/meta_bought_moltbook_openai_got_openclaw_feels/