Meta launches enterprise business to monetize its $600 billion AI bet

Meta is officially starting an enterprise business to sell its AI products and services to other companies, a move Mark Zuckerberg calls the "next major pillar of our business."

The announcement, made in a Facebook post by Zuckerberg, confirms what Meta's recent product activity has suggested for months: the company no longer wants its business to depend solely on advertising. To lead the new unit, Zuckerberg has appointed Chirantan "CJ" Desai as Meta's Chief Enterprise Platform Officer. Desai previously served as chief executive of the database company MongoDB.

The new operation will carry the name Meta Enterprise Platform. In his post, Zuckerberg laid out the rationale in unusually direct terms.

"Meta Enterprise Platform will use our strengths that few other companies have: advanced models, leading agents, large-scale infrastructure, and years of working closely with many businesses," Zuckerberg wrote. "Initially, we will focus on bringing our full technology stack, including the Muse agent, Meta Business Agent, Muse API, Muse Code, and more to businesses and developers to help them grow."

Months of groundwork

Meta has spent the last several months building the foundation for this shift. The company has released a string of AI-focused tools aimed at businesses, including business agents and a coding platform. These products now have a commercial home under the Meta Enterprise Platform banner.

There is also the question of Meta's infrastructure. Reports have suggested the company has considered leasing some of its data center capacity to other companies, though Meta has not announced specific plans to do so. Such a move would put Meta in more direct competition with cloud providers like Amazon, Microsoft, and Google, but for now that remains speculation rather than strategy.

The money problem

The commercial logic behind the new business is straightforward. Zuckerberg has said he expects Meta to spend $600 billion on AI over the next two years, with much of that going toward new data centers. Spending at that scale has raised an obvious question: how does Meta plan to make that money back?

So far, Meta's monetization experiments have targeted consumers and small business owners through AI subscriptions. The results remain unproven. Meta has yet to demonstrate that Facebook and Instagram users will pay for these offerings in large numbers. Advertising, which still generates the overwhelming majority of Meta's revenue, cannot easily absorb a $600 billion capital commitment on its own.

An enterprise business changes that equation. Companies already pay substantial sums for AI models, coding assistants, and agent platforms from vendors like OpenAI, Anthropic, and Google. Meta is betting that its combination of frontier models, agent technology, and data center capacity can capture a share of that spending — and that businesses, unlike social media users, have established budgets for exactly this kind of software.

Zuckerberg's framing of enterprise as a "pillar" puts the new unit in rare company at Meta. Advertising has been the company's only true pillar since it went public. A second revenue stream of meaningful scale would mark the most significant structural change to Meta's business model in over a decade.

What ships first

The initial product set draws on tools Meta has already built. The Muse agent, Meta Business Agent, Muse API, and Muse Code form the core of what Meta will offer businesses and developers. Zuckerberg's description — a "full technology stack" — signals that Meta intends to sell the layers together rather than as isolated products, positioning itself as an end-to-end provider rather than a model vendor alone.

Desai's appointment reinforces that ambition. His background running MongoDB, a company whose business centered on selling database infrastructure to developers and enterprises, maps directly onto the job of turning Meta's AI assets into products other companies will buy.

The stakes extend beyond Meta's own ledger. If Meta succeeds in selling AI at scale to enterprises, it adds a fourth hyperscale competitor to a market currently contested by Microsoft-backed OpenAI, Google, and Amazon-backed Anthropic — one with arguably the largest owned infrastructure base of any of them. If it fails, the questions about the $600 billion spend will only get louder.

The first test will come quickly: whether businesses sign up for Meta's stack in numbers that justify the next two years of building.