Zuckerberg says Meta's enterprise AI ambitions go beyond agents

Meta entered the enterprise AI market in June with an AI agent aimed at helping businesses with customer service and other daily operations. But on Wednesday’s second-quarter earnings call, CEO Mark Zuckerberg told investors the company’s enterprise ambitions run wider than that. “We see a large enterprise opportunity to sell to businesses, including APIs, business agents, potentially selling compute directly, and other services that we’re building for large customers,” he said, framing the push as a possible new revenue stream beyond advertising, which still drives the bulk of Meta’s business, and subscriptions, a smaller contributor.

Meta’s initial focus is serving its existing base of advertisers through AI agents that work across messaging apps, letting businesses interact with their own customers via an AI interface. Zuckerberg described the pricing model as an extension of Meta’s ad business: “Just like the ad system, effectively, we will get paid when we deliver results for those businesses,” tying the approach to the sales and partnerships Meta already has with millions of advertisers and hundreds of millions of small businesses on its platforms.

Beyond that base, Zuckerberg said Meta plans to offer external customers the coding, development, and internal productivity tools it built for its own use, opening a path to serve small and larger businesses alike. He acknowledged that selling to the enterprise is a “different muscle” than the one Meta has historically used.

On compute, Meta is weighing the revenue it could earn selling capacity — at what the company said is “a significant premium over what we paid for it” — against its own future infrastructure needs. Zuckerberg cautioned it “would be foolish” to sell all its compute for a short-term profit, describing Meta’s approach instead as a portfolio mixing long-term and short-term plans, and noting that as the company gets closer to “personal superintelligence,” it will need hardware that lets people interact with it seamlessly.

The call also covered Meta’s agentic AI ambitions for consumers, including personal AI agents and AI smart glasses, alongside a faster pace of shipping new social apps built using large language models — recent examples include apps for Marketplace sellers, Facebook Groups, and vibe-coded games, with more on the way.

Why it matters

Meta is signaling it wants enterprise AI revenue to diversify a business still dominated by advertising, while framing compute sales as a balancing act between near-term profit and the infrastructure it will need for its own longer-term AI ambitions.