Meta is framing enterprise AI as a broader business than a single customer-service agent. On Wednesday's second-quarter earnings call, CEO Mark Zuckerberg told investors that the company sees room to sell several kinds of AI products and infrastructure to businesses.
The shift matters because Meta's core business is still driven mostly by advertising, with subscriptions contributing a smaller share. Enterprise AI could give the company additional ways to make money from the tools, systems, and compute capacity it is already building.
Meta's enterprise AI plan is broader than agents
In June, Meta entered the enterprise AI market with a new AI agent for businesses. The agent is designed to help with customer service, support, and other daily operations, especially where companies already talk with customers through Meta's platforms.
Zuckerberg described a wider commercial plan during the earnings call. "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.
That list points to more than one product category. APIs could let companies connect Meta's AI capabilities into their own systems. Business agents could handle customer interactions through an AI interface. Compute sales would put part of Meta's infrastructure spending to work as a direct business service.
The common thread is that Meta is looking at enterprise AI as a set of services, not only as a chatbot-like product. The company is also presenting the opportunity as connected to work it already has underway for its own platforms and operations.
Advertisers are the first obvious customer base
Meta's initial focus is its existing advertiser base. The company already works with many businesses that use its platforms to reach customers, and Zuckerberg described AI agents as a way to extend those commercial relationships.
These agents would operate across messaging apps and elsewhere, allowing businesses to interact with their own customers through AI. For smaller businesses, that could mean using automated systems for routine customer conversations, support, and daily operational tasks.
Zuckerberg compared the model to Meta's ad system. "And, just like the ad system, effectively, we will get paid when we deliver results for those businesses," he said. He added that Meta sees this as "an extension of the sales and the partnerships that we have with many millions of advertisers and hundreds of millions of small businesses that use our platforms."
That framing is important. Meta is not starting from zero in business sales, because advertisers already spend money through its platforms. But selling enterprise AI is not identical to selling ads, and Zuckerberg acknowledged the difference directly.
Internal AI tools could become external products
Meta is also building AI tools for its own employees and development work. Zuckerberg said the company is creating coding, development, and internal productivity tools because it needs systems tuned for its own use.
Those tools may later become part of the enterprise AI opportunity. "There are other enterprise customers who I think we're increasingly going to serve, too," Zuckerberg said. He continued that once Meta has built these internal systems, it sees "a large opportunity to serve -- whether that's small businesses or larger businesses."
This suggests Meta is considering a path where internal productivity tools eventually reach outside customers. The source does not describe specific product names, pricing, or release timing for those tools. What is clear is that Meta sees them as relevant beyond its own walls.
Zuckerberg also said enterprise selling is a "different muscle" for Meta. That is a notable admission because the company has historically been strongest in consumer platforms and advertising products rather than traditional enterprise software sales.
Compute creates a short-term and long-term tradeoff
Another part of Meta's enterprise AI discussion centered on compute. The company said more than once that it has an opportunity to sell compute at "a significant premium over what we paid for it." That could create direct revenue from infrastructure capacity.
But Zuckerberg cautioned against treating compute only as a near-term profit opportunity. He said it "would be foolish" to "sell all of the compute and take a short-term profit." Instead, he described Meta's approach as a "portfolio" that balances short-term and long-term plans for its compute infrastructure.
The reason is that Meta still needs compute for its own AI ambitions. Zuckerberg linked that need to future products, saying, "As we get closer to personal superintelligence, we are . . . going to need hardware that allows you to seamlessly interact with it."
In plain terms, Meta is weighing two uses for the same resource. It can sell compute to enterprise customers, but it also needs enough capacity to build and run its own future AI systems and hardware experiences.
Agentic AI is not only for businesses
The earnings call also highlighted Meta's broader work on agentic AI. These are AI systems that can act on behalf of a person or business instead of only answering questions.
For consumers, Meta is promising "personal AI agents" and AI smartglasses that can interact with the world in front of them. For businesses, the same general direction supports agents that handle customer-facing or operational tasks.
Meta is also using AI technology, specifically large language models, to build out its suite of social apps more quickly. Recent launches mentioned in the source include an app for Marketplace sellers, another for Facebook Groups, one for vibe-coded games, and other experiments.
Zuckerberg said more products are coming. "I expect it to become a lot easier to ship new apps," he said. "So we are planning to build out more ideas and use our recommendation systems to scale them to the people who will find them interesting."
Taken together, Meta's message is that enterprise AI is one part of a wider AI strategy. Business agents are the first visible piece, but the company is also looking at APIs, internal tools, compute, consumer agents, smartglasses, and faster app development as connected parts of the same push.