Meta is building a new data center, dubbed Hyperion, which it expects will eventually deliver five gigawatts (GW) of computational power to its AI lab, CEO Mark Zuckerberg announced Monday on Threads.

 

The move highlights Meta’s latest effort to outpace OpenAI and Google in the AI race. After previously recruiting high-profile leaders—including former Scale AI CEO Alexandr Wang and former Safe Superintelligence CEO Daniel Gross to run its Meta Superintelligence Lab—Meta now appears focused on securing the vast computational resources needed to train cutting-edge AI models.

 

According to Zuckerberg, Hyperion will occupy a footprint large enough to cover most of Manhattan. Meta spokesperson Ashley Gabriel told TechCrunch that Hyperion will be built in Louisiana, likely in Richland Parish, where Meta had earlier announced a $10 billion data center. Gabriel said Meta plans to activate two GW of capacity by 2030, scaling up to five GW in the following years.

 

Additionally, Zuckerberg shared that Meta plans to bring a one GW supercluster, Prometheus, online in 2026, making Meta one of the first tech giants to directly control an AI data center of this magnitude. Prometheus will be located in New Albany, Ohio, Gabriel confirmed.

 

This AI infrastructure buildout positions Meta to better compete with OpenAI, Google DeepMind, and Anthropic in training and deploying state-of-the-art AI models. It may also help Meta attract top AI talent drawn by the promise of working with massive computing power.

 

Together, Prometheus and Hyperion will require enormous amounts of energy—enough to power millions of homes—which could strain local resources. A recent report from The New York Times noted that Meta’s data center project in Newton County, Georgia, had already led to water shortages in nearby homes.

 

Other hyperscalers face similar challenges. AI infrastructure company CoreWeave, for example, is planning a data center expansion near Dallas, Texas, projected to double the city's electricity demand, according to Bloomberg.

 

Despite these risks, tech companies remain committed to building enormous data centers to fuel their AI ambitions. Other major projects include OpenAI’s Stargate, developed in partnership with Oracle and SoftBank, and xAI’s planned Colossus supercomputer.

 

The Trump administration has largely backed the sector’s expansion. Former President Donald Trump helped OpenAI announce Stargate and has spoken publicly about the importance of strengthening America’s AI infrastructure.

 

In a Monday column in The Economist, U.S. Secretary of Energy Chris Wright argued that the U.S. should “lead the next major energy-intensive frontier: artificial intelligence.” He highlighted AI’s transformation of electricity into “the most valuable output imaginable: intelligence,” and signaled that the federal government will boost energy production from coal, nuclear, geothermal, and natural gas to meet rising demand.

 

With federal support, the AI sector appears set to absorb a significant share of the nation’s energy supply. Analysts estimate data centers could consume about 20% of America’s electricity by 2030, up sharply from 2.5% in 2022—a shift that, without rapid growth in energy generation, could further impact surrounding communities.