Перейти к содержимому
Новости

Meta Plans More Than US$50bn for 5GW Hyperion AI Campus in Louisiana

Редакция ЦОД:Главное

Data Centre Magazine — иллюстрация к материалу
Источник изображения: Data Centre Magazine

Meta is developing its Hyperion AI campus in Richland Parish, Louisiana, with planned investment exceeding US$50 billion and capacity described as growing to 5GW. The company also reports major spending on local roads, water and wastewater infrastructure, alongside contracts with local businesses.

Главное
  • Hyperion is a planned AI-campus development in Richland Parish, Louisiana, with planned investment exceeding US$50 billion.
  • The campus’s capacity is described as growing to 5GW, not as completed or operational capacity.
  • Meta reports more than US$1 billion of investment in local roads, water supply and wastewater systems.
  • More than US$1.6 billion has reportedly been contracted with local companies.
  • The project points to utility and civic infrastructure as core delivery requirements for multi-gigawatt AI campuses.

Meta is developing the Hyperion AI campus in Richland Parish, Louisiana, with planned investment exceeding US$50 billion. The company’s strategy review describes the campus’s energy capacity as growing to 5GW; this is a forward-looking capacity figure, rather than completed operational capacity.

Alongside the campus investment, Meta says it is investing more than US$1 billion in local roads, water supply and wastewater systems. It also reports that more than US$1.6 billion of work has already been contracted with local companies. The source does not disclose whether either figure is included within, or additional to, the stated total planned investment for Hyperion.

For the data-center sector, the project highlights the infrastructure profile of multi-gigawatt AI development. At this scale, campus delivery extends beyond IT and facility construction to utility networks and associated local infrastructure. If the stated parameters are realized, Hyperion could demonstrate how access to multi-gigawatt electricity supply, water and wastewater capacity, long-duration capital and local delivery capability shape the development of AI campuses.

Meta says its global portfolio comprises 33 data centers supporting Facebook, Instagram, WhatsApp and its Llama family of AI models. Separately, the company reported matching its 2024 electricity use of 18.4 million MWh with renewable energy and a PUE of 1.08. These portfolio-wide figures do not establish the energy mix or operating efficiency of the Hyperion campus specifically.

Почему это важно

The project illustrates the shift toward multi-gigawatt AI campuses, where data-center development depends not only on compute infrastructure but also on large-scale investment in utility systems and local infrastructure. If delivered as described, Hyperion could underscore how grid-scale power availability, water and wastewater readiness, long-term capital and local contractor capacity are becoming central constraints on AI infrastructure deployment.

Источники

Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.