Particle.news

Download on the App Store

Meta Commits Hundreds of Billions to Build Multi-Gigawatt AI Superclusters

Meta’s $64–72 billion capex boost for 2025 is funding rapid-build tactics under Superintelligence Labs to ready its 1 GW Prometheus cluster for 2026.

Image
Image
Image
Image

Overview

  • Meta has formalized plans to invest hundreds of billions of dollars in AI infrastructure, raising its 2025 capital expenditure target to $64–72 billion.
  • The company has centralized its AI research under Meta Superintelligence Labs led by former Scale AI CEO Alexandr Wang and ex-GitHub chief Nat Friedman.
  • Construction is underway on Prometheus in New Albany, Ohio, which is set to deliver over 1 GW of compute capacity in 2026.
  • A second “Hyperion” facility in Richland Parish, Louisiana, is designed to scale to 5 GW over several years using prefabricated power and cooling modules for rapid deployment.
  • These efforts aim to outpace rivals such as OpenAI, Google and Anthropic by pairing massive compute power with an elite talent roster.