Meta's Moltbook buy targets agentic web, not bots

News broke Tuesday morning that Meta acquired Moltbook, a social network designed for AI agents. The announcement may have puzzled some people. Why would Meta, an ad-supported company, want a platform populated by bots? After all, bots aren't exactly the target audience for brand marketers and advertisers.

Meta has not shared much detail. Its sole official statement noted that the Moltbook team is joining Meta Superintelligence Labs, which will explore “new ways for AI agents to work with people and businesses.”

Reading between the lines, this appears to be an acqui-hire. A network built for bots is hardly a natural fit for brand advertising, even though Moltbook was never purely non-human. What Meta truly sought was the talent behind it—people who enjoy brainstorming and experimenting with AI agent ecosystems. Counterintuitively, that could benefit its advertising business.

As Meta CEO Mark Zuckerberg stated last year, he envisions a future where “every business will soon have a business AI, just like they have an email address, social media account, and website.” On an agentic web—where AI systems act independently on behalf of users—AI agents could interact with one another to buy ads, make bookings, and respond to customers.

AI is also being used to create ad content and tailor it based on the viewer. AI systems could manage product pricing or generate personalized offers.

On the consumer side, agents can find the best prices and deals, manage bookings, and shop for products. In some limited cases, agents already handle checkout and payment on behalf of consumers. (Agentic commerce is still early, and these systems don't always perform as advertised. However, the market is moving quickly, and improvements are likely soon.)

Just as Facebook built the “friend graph”—a network of social connections where each person is a node—an agentic web could leverage an “agent graph,” a system that maps how agents are connected and what actions they can take on each other's behalf.

Image credit: akinbostanci (opens in a new window) / Getty Images

For an agentic web where business agents and consumer agents can collaborate, the agents must first be able to find, connect, and coordinate with each other. Just as Facebook built the “friend graph”—a network of social connections where each person is a node—an agentic web could benefit from an “agent graph,” a system that maps how different agents are connected and what actions they can perform on each other's behalf. This could cover areas like travel, online shopping, media and research, productivity tools, and more.

This is also where advertising could fit in. Currently, humans view and click on ads when something catches their interest. On an agentic web, where agents shop on behalf of users, ads might look very different. Instead of persuading a human to purchase a product, a business's agent may need to negotiate directly with a consumer's agent to close a sale.

Perhaps the consumer wants that shirt or lipstick, but only in a specific color and at a specific price. Maybe the systems become so complex that considerations extend beyond product and price—perhaps the consumer prefers to support small businesses or only shops with eco-friendly companies. Maybe they only buy items on sale or purchase generic versions if the ingredients are identical. And so on.

In that case, it's not just about connecting AI agents but also ranking products by which best suits each individual customer's needs. If Meta could capitalize on this market—AI at the orchestration layer, meaning the system decides which agents communicate and in what order—it could expand its advertising business into entirely new territory.

All of this depends on whether consumers embrace the agentic web and trust AI enough to let it act on their behalf. Yet the existence of OpenClaw, the personal AI assistant that populated Moltbook with content, suggests that at least some people are already leaning toward autonomous AI agents.

Of course, there is another possible reason Meta acquired Moltbook. The company lost the acqui-hire of OpenClaw's creator, Peter Steinberger, to rival OpenAI, so it pursued Moltbook instead—the platform that Steinberger's tool helped build. Petty? Perhaps. But it kept Meta's Superintelligence Labs in the news.

Meta's Manus news receives mixed reactions in Washington and Beijing