News · Meta's Richland Parish Update: $875M in Louisiana Contracts After One Year of Hyperion Construction
Meta's Richland Parish Update: $875M in Louisiana Contracts After One Year of Hyperion Construction
A year into building its largest multi-gigawatt AI training cluster, Meta's progress report reads less like an engineering milestone and more like a regional economic ledger.
What Meta reported at the one-year mark
Meta broke ground on its Richland Parish Data Center in December 2024, and this December 19 update is the twelve-month accounting. The headline figure: more than $875 million contracted with Louisiana businesses over the past year.
The site is described as a four-million-square-foot campus that will house Hyperion, which Meta calls its largest multi-gigawatt AI training cluster. Construction is led by three general contractors — Turner Construction, DPR Construction, and Mortensen Construction.
Meta reports working with more than 160 Louisiana businesses, 84% of them local to Northeast Louisiana, covering electric, paving, and utility work, plus local food vendors on site. The company says it has supported 3,700 construction workers and expects a peak workforce of 5,000 by June 2026, with more than 500 permanent operational jobs once the facility is complete.
The energy and water figures Meta chose to publish
The most concrete numbers in the release concern the grid. Meta says it worked with Entergy so that electricity payments for the data center would cut grid upgrade customer costs and storm charges by about 10%, which it frames as $650 million in customer savings over 15 years. It is also contributing $15 million to Entergy's customer assistance program for low-income customers.
On water, Meta commits to returning 100% of the data center's water consumption to the Boeuf, Tensas, and Lower Mississippi watersheds, working with Resource Environmental Solutions and Ducks Unlimited on five wetland restoration projects.
These are the line items that matter for a facility drawing multiple gigawatts. A multi-gigawatt training cluster is an enormous, sustained load, and the release is careful to state that Meta pays for the energy and infrastructure it uses so other customers' bills don't rise. That framing is itself a signal of where public scrutiny of AI infrastructure now sits.
The community spending that surrounds the compute
Beyond the core build, Meta lists more than $300 million for local roads, water, and wastewater systems, plus $300,000 to the Richland Revitalization Board for parks and landmark restoration across Rayville, Delhi, and Mangham.
Smaller items round it out: senior services at the Richland Voluntary Council on Aging, upgrades at the Northeast Louisiana Veteran's Cemetery, a robotics program at Delhi Charter School, and mixed reality workforce development at Richland Parish Schools. A Data Center Community Action Grant Program will name recipients in spring 2026.
Meta's decision to invest here has done more than bring a world-class project to Louisiana — it is activating entire ecosystems. Over the past year, we've seen local contractors, technology firms, and service providers across the state expand their teams and capabilities to support this project.Montana Labs
Why the physical footprint of frontier training is now the story
What's notable about this announcement is what it does not contain. There is no model, no benchmark, no capability claim beyond a single phrase about accelerating progress toward personal superintelligence. The entire document is about concrete, watersheds, kilowatts, and payroll.
For anyone building on top of these systems, that is the underappreciated reality: the interface a user touches sits on top of a four-million-square-foot campus, 5,000 construction workers, and a 15-year utility settlement negotiated with a state economic development office. The compute behind an AI product increasingly has a street address, a workforce, and a local politics.
The specific implication of the Richland Parish update is that Meta is now treating the social license to build training capacity as a deliverable in its own right — reporting contractor spend, utility protections, and water returns with the same care it once reserved for model releases. Hyperion isn't operational yet, but the infrastructure balance sheet is already public.
Find this story relevant to you?
Contact us to find a unique solution
Need an AI engineering partner that can actually build?
We help businesses integrate AI, build AI-powered products, automate high-value workflows, and modernize the software systems behind them.
Related reading
More analysis around product delivery, operational AI, and the systems work that makes deployment hold up in reality.