Overview

• Meta has pledged to invest hundreds of billions of dollars into compute infrastructure aimed at building AI superintelligence
• Some of the new AI data centers are being erected in temporary tent facilities to meet ambitious deployment timelines
• Prometheus, Meta’s first supercluster, will deliver one gigawatt of capacity when it goes online in New Albany, Ohio, in 2026
• Hyperion, slated for Richland Parish, Louisiana, will scale up to five gigawatts over several years with additional “titan” clusters planned
• Meta Superintelligence Labs has been formed under Alexandr Wang and Nat Friedman following aggressive talent recruiting to support the infrastructure rollout