Meta Enterprise Platform: Zuckerberg Turns the Social Media Giant Into an A.I. Neocloud

Set Trending Topics as a preferred source on Google.

Anyone pouring $130 billion into data centers will eventually want to send out invoices for them. Meta has just announced a new business unit called Meta Enterprise Platform, through which the company will for the first time sell its A.I. models, agents and infrastructure as a bundle to businesses and developers. The division is led by Chirantan “CJ” Desai, until now the CEO of MongoDB, who joins as Chief Enterprise Platform Officer and reports directly to Mark Zuckerberg. With that, the parent company of Facebook and Instagram is entering a business that has so far been the domain of hyperscalers like Microsoft, Google and Amazon as well as specialized A.I. cloud providers.

“We believe superintelligence will create significant new opportunities for all people and businesses,” Mr. Zuckerberg writes in the announcement. “Today we are starting the next major pillar of our business, Meta Enterprise Platform, to help businesses use AI to grow and transform in new ways as well.” The new unit will “use our strengths that few other companies have: advanced models, leading agents, large-scale infrastructure, and years of working closely with many businesses.”

What Meta Wants to Sell to Businesses

To begin with, Meta is bringing “our full technology stack” to businesses and developers, according to Mr. Zuckerberg. Specifically named are:

• Muse agent: The personal A.I. agent Meta launched a few weeks ago for consumers is to take on tasks inside companies as well.
• Meta Business Agent: The agents for WhatsApp, Instagram and Messenger introduced over the summer answer customer questions, book appointments and close sales.
• Muse API: Programmatic access to Meta’s models, letting companies build their own applications.
• Muse Code: The recently launched coding agent that takes on Anthropic’s Claude Code and OpenAI’s Codex with aggressive pricing.

Meta has not said how the offering will be priced or when it will be available outside the United States. Mr. Desai, who previously led product and engineering at Cloudflare and spent nearly eight years at ServiceNow, most recently as President and COO, puts the goal this way: “Our goal is to make Meta the place enterprises come to scale their businesses.” As with Muse, he says, security and privacy are built into the enterprise products from the outset.

From Ad Business to Neocloud

The real core of the announcement is strategic. By its own forecast, Meta is spending between $130 billion and $145 billion (about 111 billion to 124 billion euros) on capital expenditures this year alone, most of it on A.I. data centers. Until now, the company built that infrastructure exclusively for itself: for the recommendation algorithms of Facebook and Instagram, the ad system and its own A.I. models. The Enterprise Platform changes that. Meta is now selling access to models, agents and, down the line, raw computing power, which is exactly what so-called neoclouds like CoreWeave, Nebius or Lambda do.

Mr. Zuckerberg had already sketched out the direction at the last quarterly report. “We see a large enterprise opportunity to sell to businesses, including APIs, business agents, potentially selling compute directly, and other services,” he told analysts. Compute, he said, could be sold at “a significant premium over what we paid for it.” Still, it would be “foolish to sell all of the compute and take a short-term profit.” Back in the spring, he had described a cloud business of Meta’s own as “definitely on the table.” Selling to enterprises, Mr. Zuckerberg conceded, requires “a different muscle” than the ad business. That is precisely why Mr. Desai was brought in.

The pressure from Wall Street is hard to miss. After the last quarterly report, in which Meta raised its capex forecast once again and net income fell despite a 28 percent jump in revenue, the stock dropped sharply. The success of the Muse agent in recent weeks has lifted the mood, but investors want to see how the multibillion-dollar bet turns into revenue beyond advertising. The Enterprise Platform is the answer: subscriptions, API fees and compute contracts instead of ads alone.

Up Against Amazon, Microsoft, Google and OpenAI

Meta is entering a fiercely contested market. Amazon Web Services, Microsoft Azure and Google Cloud have offered computing power, models and developer tools from a single source for years, on top of productivity suites like Microsoft 365 and Google Workspace that bring A.I. agents straight into everyday work. OpenAI and Anthropic sell their models directly to enterprises and have loyal developer communities around Codex and Claude Code. Neoclouds like CoreWeave and Nebius, in turn, specialize in renting out pure GPU capacity and are growing rapidly. And low-cost open-weight models from China are adding to the price pressure.

Meta’s argument against all of that: the company already has relationships with hundreds of millions of businesses that advertise or serve customers on Facebook, Instagram and WhatsApp. Anyone already buying ads there is meant to source agents, models and computing power from the same place in the future. Whether that approach works will depend on whether Meta Enterprise Platform wins over companies beyond its own advertising base that currently buy from the hyperscalers. As Amazon recently showed with the Muse agent, the incumbents will not make it easy for Meta.