Meta reported $60.8 billion in Q2 2026 revenue, up 28% year-on-year, and Mark Zuckerberg used the earnings call on July 29 to make something clear: the company’s ambitions now extend well beyond consumer social media.

“We see a large enterprise opportunity to sell to businesses,” Zuckerberg told analysts, listing “APIs, business agents, potentially selling compute directly, and other services that we’re building for large customers.”

This is a meaningful shift. Meta built one of the world’s largest technology businesses on advertising. The new thesis is that the same AI infrastructure powering its ad targeting, recommendation systems, and consumer features can be sold directly to enterprises, competing in the same market as OpenAI, Anthropic, Google, and Microsoft.

What Meta Is Actually Selling

The enterprise strategy has a few distinct pieces.

Business agents for customer-facing and operations workflows, initially targeting WhatsApp, Messenger, and Instagram as the delivery channels. Meta entered this market in June 2026 with an AI agent product aimed at customer service and support. The Q2 call expanded that narrative considerably.

APIs and models that give enterprises access to Meta’s Llama family directly, competing with the commercial API offerings from OpenAI and Anthropic. Meta has positioned Llama as open-weight, which has made it popular with businesses that want to run models on their own infrastructure. The enterprise API play is about monetising that trust with a supported commercial product.

Compute sold directly to large customers, following a path similar to what Amazon, Google, and Microsoft have been doing for decades. Meta has invested heavily in its own infrastructure and is now exploring whether it can earn a return on that investment beyond using it internally.

Productivity tools for internal enterprise use, described as “coding and productivity tools for larger businesses.” This brings Meta into direct competition with Microsoft Copilot and Google Workspace AI.

The Spending Question

Zuckerberg’s enterprise pivot comes as Meta is spending at a scale that is hard to overstate. The company guided 2026 capital expenditures to $130 billion to $145 billion, nearly double its 2025 figure of $72.2 billion. Almost all of that is AI infrastructure: data centres, custom chips, power infrastructure, and the talent to run it.

At a July 2 internal town hall, Zuckerberg acknowledged that Meta’s AI development “hasn’t really accelerated in the way that we expected.” The reorganisation that shifted thousands of employees into new AI teams has been messier than planned, and the financial returns are still being built.

The enterprise push is partly an attempt to find a return path for that spending. Advertising alone, even at 28% growth, may not be enough to justify $130-145 billion in annual capital investment. Enterprise software and AI services, historically high-margin businesses, represent a different financial profile if Meta can execute on them.

What This Means for Business

Meta’s entrance into enterprise AI is a market signal worth paying attention to.

More competition means better choices for buyers. Enterprise AI was already getting crowded, with OpenAI Presence, Anthropic Claude for Enterprise, Google Vertex AI agents, and Microsoft Copilot all competing for the same buyer. Meta adds another well-resourced competitor with a distribution advantage: billions of people already use WhatsApp and Messenger for business communication in many markets.

Open-weight models remain a credible alternative. Meta’s Llama models give businesses a way to run capable AI on their own infrastructure without ongoing API costs or data exposure concerns. The Llama commercial API plays into that existing trust.

The enterprise AI market is still being defined. The fact that Meta, a consumer-first company, is now positioning itself as an enterprise AI vendor illustrates how early-stage this market still is. The winners five years from now may not be obvious from where the starting positions sit today.

Meta’s scale creates pressure on pricing. When a company with $60 billion in quarterly revenue and its own chip infrastructure enters a market, it can afford to compete aggressively on price. Enterprise AI pricing that looks stable now may face significant downward pressure as Meta’s enterprise products mature.

For business leaders building their AI strategy, the actionable takeaway is not to bet heavily on any single vendor. The enterprise AI landscape is shifting fast enough that platform lock-in is a genuine risk. Building internal capability, particularly the ability to understand, evaluate, and switch AI tools, is more valuable than any particular tool choice.

The skills to make that judgment are exactly what Enterprise DNA builds. If your team is trying to keep pace with decisions like this, our learning platform gives them the foundation to evaluate AI tools critically, not just use them.