Meta Unveils Multi-Gigawatt AI Data Centers Plan

Meta CEO Mark Zuckerberg has announced plans to invest hundreds of billions of dollars in building multi-gigawatt data centers across the US to advance artificial general intelligence, including the Prometheus facility in New Albany, Ohio, set to launch in 2026, and the larger Hyperion cluster in Louisiana, which could scale up to 5 gigawatts. These efforts are central to Meta's new Superintelligence Labs division and include the recruitment of prominent AI leaders such as Scale AI's Alexandr Wang as chief AI officer. Meta is also developing proprietary AI chips to reduce reliance on Nvidia and further compete with OpenAI, Google DeepMind, and Anthropic. The scale of these facilities is expected to rival or surpass those of competitors. However, the massive energy and water requirements have raised concerns about local resource strains. Despite the aggressive infrastructure investment, analysts warn that significant revenue may not be realized for several years.