Meta Bets Big on AGI with Massive AI Data Centers Prometheus and Hyperion
Meta is making a bold move to lead the global race for artificial general intelligence (AGI). CEO Mark Zuckerberg recently announced the development of two massive AI data centers—Prometheus and Hyperion—designed to deliver unprecedented computational power.
In a post on Threads, Zuckerberg revealed that the upcoming Hyperion facility, located in Richland Parish, Louisiana, will eventually scale up to 5 gigawatts (GW) of capacity—an area footprint comparable to Manhattan. The company plans to bring 2 GW online by 2030, with full expansion in the years that follow. Hyperion will support Meta’s newly launched Superintelligence Labs, a division dedicated to building next-generation AI systems.
Joining Hyperion is Prometheus, a 1 GW AI supercluster based in New Albany, Ohio, set to launch in 2026. It will be one of the first multi-gigawatt AI data centers fully owned and operated by a tech company.
Together, these facilities reflect Meta’s aggressive strategy to train and deploy frontier AI models at scale, putting it in direct competition with OpenAI, Google DeepMind, and Anthropic.
Meta has committed $64–$72 billion in capital expenditures by 2025, much of it allocated to AI infrastructure. With $165 billion in advertising revenue last year, Zuckerberg emphasized the company's ability to self-fund these ambitious projects.
Attracting elite AI talent is another key part of Meta’s play. The company has recruited Alexandr Wang, former CEO of Scale AI, and Nat Friedman, ex-GitHub CEO, to lead Superintelligence Labs. Reports suggest Meta is offering multimillion-dollar compensation packages—some topping $100 million—to lure top researchers away from competitors.
However, Meta’s rapid AI buildout has raised environmental concerns. A New York Times investigation revealed that a Meta data center in Georgia depleted local water resources, leaving some homes without running water. Experts warn that data centers could consume up to 20% of the U.S.’s electricity by 2030, a steep rise from 2.5% in 2022. These concerns are explored in detail in Karen Hao’s book, "Empire of AI," particularly in Chapter 12, which examines the sector’s environmental impact.
The U.S. government has expressed strong support for expanding AI infrastructure. In a recent Economist column, Energy Secretary Chris Wright called on the U.S. to lead this new energy-intensive frontier, advocating for increased energy production from coal, nuclear, geothermal, and natural gas to support AI growth. The Trump administration has also backed efforts like OpenAI’s Stargate project.
With Prometheus and Hyperion, Meta is positioning itself at the forefront of the AGI race—aiming to integrate superintelligent systems across its platforms, from social media and smart glasses to virtual reality and the metaverse. If successful, these initiatives could reshape human–technology interaction and redefine the future of intelligence.