Meta's Louisiana data center scales to 5GW in $50B AI bet
Meta is more than doubling the planned compute capacity of its Louisiana data center to 5 gigawatts, pushing the site's total investment past $50 billion as the company races to secure power for large language model training.
Meta said on Monday its Hyperion data center in Richland Parish, Louisiana, will scale to 5 gigawatts of compute capacity, pushing total investment in the project beyond $50 billion. The facility was previously slated to deliver just over 2 gigawatts of power dedicated to training large language models.
The 150% increase in planned capacity highlights the staggering and evolving capital requirements of the artificial intelligence buildout. Like its Big Tech peers, Meta is pouring billions into computing infrastructure as demand for AI processing continues to vastly outstrip available supply. The company has committed to investing $600 billion in U.S. infrastructure and jobs over the next three years to support CEO Mark Zuckerberg's aggressive push into AI agent technologies.
Construction on the Louisiana site began in December 2024. Since breaking ground, Meta has directed more than $1.6 billion in contracts to local Louisiana businesses. The company confirmed it will also invest over $1 billion in surrounding public infrastructure, including necessary upgrades to roads, water, and wastewater systems to support the massive facility.
Such colossal projects carry significant structural risks, particularly regarding power purchase agreements and utility financing. Environmental and consumer groups are actively pushing back against the energy-intensive expansion. Earlier this year, the U.S. environmental law group Earthjustice requested an investigation into the project's financing structure.
Earthjustice argued the arrangement could unfairly shift project costs onto utility ratepayers if Meta were to walk away before the utility recouped its investment. Regulators denied that investigation request. The $50 billion valuation matches a figure cited last year by U.S. President Donald Trump.
For market participants, the Hyperion expansion serves as a stark reminder of the exponential cost curves now defining the current AI cycle. Securing adequate, reliable power has emerged as the primary bottleneck for tech giants. This constraint is forcing companies to dramatically scale their initial project estimates just to maintain their competitive positions in the large language model market.
