Mark Zuckerberg told investors on Meta's second-quarter earnings call Wednesday that the company's enterprise AI ambitions extend well past the customer-service agent it launched in June 2026, laying out a pay-for-results business that would sell APIs, autonomous agents, raw compute and internal productivity tools to outside customers. The pitch reframes Meta, historically an ad-driven consumer platform, as a full-stack enterprise AI vendor competing directly with the incumbents that dominate that market today.
The addressable base Zuckerberg described is unusually large by enterprise-software standards: hundreds of millions of small businesses and many millions of advertisers already transacting on Meta's platforms. That existing distribution is the wedge — Meta plans to bolt AI agents onto messaging apps and other surfaces those businesses already use, then charge on outcomes rather than seats or tokens.
The revenue model matters as much as the product. Meta's ad system is priced on delivered results, and Zuckerberg said the enterprise AI business will follow the same structure — a departure from the per-token or per-seat pricing that defines the current enterprise AI market. If it works, it aligns Meta's incentive with customer ROI in a way subscription-based rivals cannot easily match.
Key facts
- 01Meta is pitching enterprises on APIs, business agents, direct compute resale and productivity tools beyond its June 2026 customer-service agent launch.
- 02The initial addressable base is hundreds of millions of small businesses and many millions of advertisers already on Meta's platforms.
- 03Zuckerberg said Meta can currently resell compute at 'a significant premium over what we paid for it' but won't liquidate its stack for short-term profit.
- 04Meta plans to externalize internal coding and productivity tools it built for itself, moving upmarket from SMB advertisers to larger enterprise customers.
- 05Recent LLM-built app launches include tools for Marketplace sellers, Facebook Groups and vibe-coded games, with more on the roadmap.
Zuckerberg framed the move as a natural extension of Meta's existing commercial relationships rather than a cold start into a new market. He acknowledged that selling to enterprises is a "different muscle" than the one Meta has flexed for the past two decades, a rare note of hedging from a company that has historically bulldozed into new categories.
“And, just like the ad system, effectively, we will get paid when we deliver results for those businesses.”— Mark Zuckerberg, Meta CEO
The compute-resale angle is the more surprising line. Meta said it currently has the opportunity to sell compute to enterprise customers at "a significant premium over what we paid for it," but Zuckerberg cautioned it "would be foolish" to "sell all of the compute and take a short-term profit." The company is treating its infrastructure as a portfolio, balancing what it monetizes externally against what it reserves for its own model training and product roadmap.
That roadmap increasingly hinges on what Zuckerberg calls personal superintelligence, and on hardware to deliver it. Meta is also pushing agentic AI to consumers through personal AI agents and AI smart glasses that can perceive the wearer's environment — a consumer bet we covered last week when Zuckerberg first outlined personal AI agents as Meta's next revenue engine.
The enterprise pitch is a parallel bet on the same underlying stack. By productizing the coding and internal productivity tools Meta built for its own engineers, the company is following the same playbook Amazon used with AWS: dogfood the tooling, then rent it out. Zuckerberg said Meta feels it now has tools worth serving to outside customers, whether small businesses or larger enterprises.
“As we get closer to personal superintelligence, we are . . . going to need hardware that allows you to seamlessly interact with it.”— Mark Zuckerberg, Meta CEO
App velocity is the other tell. Meta has recently shipped LLM-built apps for Marketplace sellers, Facebook Groups and vibe-coded games, with more planned. Zuckerberg said he expects it to become "a lot easier to ship new apps," and that Meta will lean on its recommendation systems to distribute them to the users most likely to engage — a distribution advantage no pure enterprise AI vendor has.
The skeptic's case is the one Zuckerberg himself acknowledged: Meta has no enterprise sales motion at scale, no track record servicing Fortune 500 IT departments, and no established relationships with the CIOs who write the checks. Microsoft, Google and Amazon have decades of enterprise trust and existing procurement contracts. Selling business agents to a corner-store advertiser is not the same as selling compute and APIs to a regulated bank, and the sales cycle, support burden and compliance requirements are categorically different.
The strategic read is that Meta is trying to convert an advertising monopoly into an AI-infrastructure business before the ad market's growth flattens under agentic competition. If AI agents intermediate consumer purchases at scale, Meta's ad take rate is at risk — and outcome-priced enterprise agents on top of Meta's compute stack are a hedge against that scenario, not just a growth bet. Whether the company can build enterprise muscle fast enough to matter is the open question, but the framing tells investors that Meta no longer sees itself as purely an ad business.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
