Meta has hired Dave Brown, a senior executive from Amazon Web Services, to lead its data center expansion and build something internally called Meta Compute. Brown spent nearly 20 years at AWS before the move and will report directly to Meta’s head of infrastructure when he joins in the coming weeks.
Meta expects to spend between $125 billion and $145 billion on capital expenditure in 2026, with much of that going toward data centers, networking gear, and compute capacity for AI training and inference. The Louisiana AI data center project is scaling from 2 to 5 gigawatts, with investment exceeding $50 billion.
What Is Meta Compute?
Meta Compute appears to be an initiative to rent AI infrastructure to outside customers—essentially Meta’s version of AWS or Microsoft Azure. The company has built the capabilities internally to run AI, but selling access to that compute could turn infrastructure into a business line.
This matters because Meta has historically been a consumer company. Opening cloud compute to enterprises means entering terrain where AWS and Azure have dominated for over a decade.
The Infrastructure Arms Race
Every major AI company is racing to build data center capacity. OpenAI, Anthropic, Google, and Microsoft are all spending tens of billions. Meta’s hiring of a cloud infrastructure veteran signals the company is serious about not just building enough compute for its own models, but for selling it.
Brown’s 20 years at AWS mean he understands what it takes to run reliable, scalable cloud infrastructure at global scale. He’s not starting from scratch.
Zuckerberg’s Bet
Mark Zuckerberg has committed Meta to massive capital spending despite criticism from investors about the pace of returns. Bringing in Brown from AWS signals confidence that this spending will eventually pay off—not just for running Meta’s own AI, but through commercializing it.
Whether Meta can compete with AWS in enterprise infrastructure is an open question. Having the right person to lead the effort is a necessary first step.




