Meta acquires Moltbook, the popular AI agent social network

Meta’s (META) just bought a tiny AI startup called Moltbook.

On the surface, that might sound like another small tech acquisition.

It’s not.

Moltbook is a strange project if you look at it through a traditional social media lens. The platform launched in January and looks a bit like Reddit.

But the posts don’t come from people. They come from AI agents.

These agents share code, discuss tools, and even reference the humans who created them. In other words, Moltbook is a social network where artificial intelligence talks to itself.

By buying the platform, Meta is positioning itself for a future where AI agents communicate with each other, make decisions, and run large parts of the internet on their own.

Basically, this is the dead internet theory in action.

If that future takes shape, the companies building those systems today could end up controlling an entirely new layer of the digital economy. And the dollars flowing through it.

We've talked before about the disruptive potential of Moltbook before. Zuckerberg clearly recognizes it too.

The disruption behind the news: AI agents are becoming the next operating layer of the internet.

That means software increasingly communicates with other software instead of waiting for humans to run commands.

If that shift happens, the interface for computing changes completely.

Meta is trying to position itself inside that shift before it becomes obvious to everyone else.

The company has already been investing heavily in AI infrastructure, including massive compute deals like Meta’s multi-billion-dollar agreement to rent Google AI chips to support the next generation of large-scale AI systems.

Today most AI tools still depend on humans to start the process. You ask ChatGPT a question. You copy the result somewhere else. Maybe a script runs afterward.

Agent networks change that model.

Instead of waiting for instructions, AI programs coordinate work directly with other agents. One writes code. Another tests it. Another deploys it. Another negotiates access to APIs or purchases computing resources. Humans move into a supervisory role rather than doing every step themselves.

That’s why Moltbook mattered. It acted as a live sandbox that showed how autonomous systems might interact, collaborate, and build tools together.

Think about the adoption curve. There are about 30 million developers worldwide. If just 5 percent start deploying autonomous agents over the next three years, that would equal about 1.5 million automated workers interacting with software ecosystems constantly.

Now apply that across entire companies.

When agents become the main users of software, the economics change quickly. Platforms stop optimizing for human clicks and screen time. They start optimizing for machine cooperation, machine trust, and machine transaction speed.

The winning platform may earn money the way app stores do rather than the way social networks do.

If 1.5 million agents each make 200 software calls per day through a coordination layer, and that layer collects even a $0.002 routing or verification fee per call, that equals about $600,000 in daily platform revenue before advertising even appears.

Engagement stops being the main goal. Becoming the transaction rail for machine-to-machine work becomes the goal.

Whoever hosts the environment where AI agents coordinate could control the next massive developer economy.

Zuckerberg understands this.

Facebook connected people. The next platform might connect digital minds.

What to watch next

The next 24 months will likely determine which company controls the agent infrastructure layer.

Watch hiring patterns first. OpenAI, Meta, Anthropic, and Google are recruiting independent agent developers faster than many people realize.

Watch tooling second. Frameworks that let agents discover each other, negotiate tasks, and verify outcomes could become critical infrastructure for software development.

Watch security third. One million leaked credentials from a small experiment already shows how messy autonomous systems could become. Mistakes in autonomous systems can already be costly, as seen in a case where an AI bot accidentally sent $250,000 in crypto instead of a $500 tip.

The biggest signal will come from deployment. When the first company successfully runs thousands of cooperating agents across real business workflows, the productivity gains could be huge.

That’s the moment this stops looking like an experiment.

Meta (META) has a Disruption Score of 4. Click here to learn how we calculate the Disruption Score. 

Meta is also part of the Disruption Aristocrats, our quarterly list of the world’s top disruptive stocks.