Meta is hiring AWS's Dave Brown and talking to Anthropic as its cloud push becomes real

Meta is recruiting Dave Brown, one of Amazon Web Services' most senior vice presidents, to lead its data-center buildout, and the company is in early talks to sell spare computing capacity to Anthropic. Together, the moves confirm that Mark Zuckerberg's cloud ambitions, which he called 'definitely on the table' at Meta's May shareholder meeting, are becoming concrete hires, deals, and organizational decisions.

Brown will report to Meta's head of infrastructure and is expected to join in the coming weeks, according to the Wall Street Journal. He brings nearly two decades of cloud infrastructure experience from AWS. His arrival gives Meta a credible operational leader for a business it has not yet formally launched but is actively building toward.

From internal asset to external product

The strategic logic is straightforward. Meta has committed to spending as much as $145 billion in capital expenditures in 2026, the high end of guidance the company set in April, according to CNBC. It funded part of that with a $25 billion bond sale timed alongside its first-quarter earnings report. That level of infrastructure investment was always going to exceed what Meta's own applications could absorb, and Zuckerberg acknowledged as much publicly: companies were approaching Meta 'almost every week' seeking access to its AI models or spare compute, he said at the shareholder meeting.

The Anthropic talks, reported by the Wall Street Journal, represent the first named potential enterprise customer. The conversations are early and no terms have been agreed, the Journal noted, but the fact that Meta is at the table with one of the highest-profile AI labs in the industry illustrates the kind of B2B customer it is targeting. For enterprise infrastructure buyers, that framing matters: Meta would be positioning itself as a capacity provider to organizations running large AI workloads, not just a platform for consumer applications.

The margin trade-off enterprise teams should understand

Cloud infrastructure is lucrative in absolute terms for Amazon, Microsoft, and Google, but it carries structurally lower margins than Meta's core advertising business. That asymmetry is the central concern analysts have raised about the move. Paul Meeks, head of technology research at Freedom Capital Markets, told CNBC that the cloud push appears partly designed to address investor skepticism about whether Meta will ever earn a proportionate return on its massive capex. He noted that Meta has so far built capacity almost entirely for its own use and has not yet monetized AI applications directly.

The distinction matters for operational buyers evaluating Meta as a potential vendor. A cloud business built on surplus capacity is a different commercial proposition than one architected from the ground up for external customers, the way AWS and Azure were. Reliability guarantees, support tiers, and service-level agreements are all still to be defined. Karan Ramchandani, managing director at advisory firm Post Oak Group, told CNBC that the B2B compute market is a natural fit for Meta given its infrastructure scale, but the product details have not been settled. Meta is still debating whether to offer hosted AI model access or raw compute capacity, CNBC reported.

What the talent signal tells operators

The Brown hire is itself an informational signal for the market. AWS executives at his seniority level have deep experience designing the reliability, redundancy, and multi-tenant architecture that enterprise customers expect from a cloud provider. His mandate at Meta, according to the Wall Street Journal, centers on the data-center buildout rather than product management or sales, which suggests the immediate priority is infrastructure quality and scale before external commercialization.

For IT and procurement leaders, that sequencing sets a rough timeline. A vendor still hiring its infrastructure leadership and still in pre-term talks with its first external customer is likely at least a year away from offering enterprise contracts at scale. That window is still worth monitoring: Meta's AI model ecosystem, particularly the open-weight Llama family, already has significant enterprise adoption, and a cloud offering that bundles compute with model access could be differentiated from what AWS, Azure, and Google Cloud currently provide.

Meta's stock jumped 9% on the first trading day of the third quarter when CNBC reported the cloud plans, its sharpest single-day rally in more than five months. The stock had lost close to a quarter of its value over the four preceding quarters. The market's reaction reflects how significant the revenue diversification story is, but for operators, the more immediate question is when Meta's cloud offer becomes something they can actually evaluate, contract, and run workloads on.