Meta is taking a major step in the AI arms race by building one of the world’s most powerful data centers, known as Hyperion, which is expected to provide up to five gigawatts (GW) of computing power for AI development. CEO Mark Zuckerberg revealed the plan in a post on Threads, noting that the facility will supply Meta’s growing AI lab and help position the company alongside top competitors like OpenAI and Google.
The new facility will support Meta’s push to develop frontier AI models, a goal that has already seen the company recruit major talent, including former Scale AI CEO Alexandr Wang and Safe Superintelligence co-founder Daniel Gross. According to Zuckerberg, “Hyperion’s footprint will be large enough to cover most of Manhattan.”
Meta spokesperson Ashley Gabriel confirmed the data center will be located in Louisiana, likely in Richland Parish, where Meta previously announced a $10 billion investment in data infrastructure. Gabriel also stated that Meta will activate two GW of Hyperion’s capacity by 2030, eventually scaling up to five GW.
In addition to Hyperion, Meta plans to launch a separate 1 GW supercomputer cluster, Prometheus, in 2026. “Prometheus is located in New Albany, Ohio,” said Gabriel, adding that it will make Meta one of the first tech firms to control a dedicated AI data center of such scale.
These developments are expected to boost Meta’s ability to train large-scale AI models and could help attract more AI researchers interested in high-performance infrastructure. However, the energy demands of these facilities are raising concerns.
The power needed for Hyperion and Prometheus is comparable to the electricity consumption of millions of households. A previous Meta data center project in Georgia reportedly left some residents without water, according to The New York Times. Similar issues may arise as other companies, like CoreWeave, expand their AI data operations across the U.S.
Government backing appears strong, particularly from former President Donald Trump, who has publicly supported AI infrastructure efforts like OpenAI’s Stargate project. U.S. Secretary of Energy Chris Wright echoed these sentiments in The Economist, calling for the country to lead in AI-powered innovation: “AI transforms electricity into the most valuable output imaginable: intelligence.”
Experts warn that data centers could consume up to 20% of U.S. electricity by 2030, up from just 2.5% in 2022 — a shift that may strain communities if energy production fails to keep pace.
