Meta’s $27 billion AI data center is causing chaos in small town Louisiana

On a recent morning, the AI boom in Richland Parish, a rural county in northeast Louisiana, could be measured in tacos.

Recommended Video

Tim and Lindsey Allen were preparing over 1,600 of them with names like “Divine Swine” (smoked pork), “Righteous Rooster” (braised chicken), and “Golden Calf” (brisket), for construction workers building Meta’s massive 2,250-acre, 4-million-square-foot AI data center, Hyperion. It’s a catering order that would have been unthinkable here just a year ago.

The Allens, parents of five, had long joked about starting a taco joint called Holy Tacos. (Tim is a church administrator and children’s pastor at the First Baptist Church in the small Richland Parish town of Rayville.) When Meta announced in December 2024 that it was investing in a $10 billion facility in Richland Parish, its largest data center to date, they saw a rare opening. Thousands of construction workers, they’d heard, would soon descend on the site—an unheard-of customer base for this otherwise rural, economically depressed community.

At first, the plan was to park a taco truck at the site. But when Allen learned that the vehicle he had invested in wouldn’t be allowed inside the construction zone, he rented a small vacant building in Rayville, pulled the truck inside, and turned it into a makeshift restaurant serving “food worth praising.”

The risk paid off. Workers coming off 12-hour shifts in safety gear began to stop by for quick, to-go meals. And as Meta’s construction ramped up, the Allens landed recurring catering work with Mortenson, one of the project’s three major contractors.

“Just this month, we picked up about 10 caterings,” Allen said. Without Meta, he added, the family likely wouldn’t have taken the leap. “It’s been a huge blessing for us.”

For Allen and many other local business owners, Meta’s arrival has brought new customers, contracts, and long-deferred opportunities in a parish that had been losing population and jobs for decades.

The outcome has been very different for Katie and Logan Stewart. The couple—a nurse and a former farmer in their mid-thirties with two children—invested more than $40,000 of their life savings into Opal’s Orange Food Truck after seeing construction workers post on Facebook asking for food options near the Meta site.

It was a high-stakes bet. Logan had recently stepped away from farming land near the project, and the food truck—serving burgers, chicken, gumbo, and rice and beans—seemed like a way to build a new livelihood without leaving the community.

“I think we were like the second or third truck out there,” Katie said. “When we first started, we were doing 100 to 120 orders a day.”

But the momentum didn’t last. When one of the project’s main contractors, DPR, brought in an out-of-state catering company to feed workers on-site, much of the foot traffic Opal’s had counted on disappeared. Workers no longer needed to leave the grounds for lunch.

“You talk about supporting the local community, but then you outsource the work,” Katie said of DPR’s catering decision. “It felt like a slap in the face.”

The Stewarts couldn’t afford the $1,500 to $2,500 monthly fees charged by two new food truck parks located right across the street from the main Meta entrances, so they parked their truck on a friend’s land a short drive away. Orders fell to fewer than 40 a day. In recent weeks, foot traffic picked up after a local independent journalist wrote about Opal’s, and the Stewarts recently landed a catering job with Meta. They say they’re determined to adapt. “We’re planning on sticking it out and adjusting where we need to,” Katie said.

Stories like those of the Allens and Logans are playing out across the country, as companies such as Meta, Google, and Amazon—alongside fast-growing AI startups like OpenAI and Anthropic—embark upon an unprecedented AI data center spending spree. Collectively, they are projected to invest roughly $630 billion to $700 billion in 2026 alone, a 62% jump from 2025, with total AI-related data-center capital expenditures expected to reach $5.2 trillion by 2030, driven largely by GPUs and energy infrastructure. These mega-scale projects—built to power the AI boom and bolster the U.S. race with China for technological dominance—are helping to grow the U.S. economy, and are being welcomed with open arms by local officials eager for a piece of the economic development these projects promise. A rising tide, they reason, can lift many ships.

And indeed, there’s historical precedent for this optimism: The frenzied construction of massive AI data centers across the country echoes earlier American booms—from the California Gold Rush to the early oil fields of Texas—when fortunes were made by those in the right place at the right time, selling equipment, food, and shelter to the pioneers of new industries. In those eras, local economies thrived on demand for tools, timber, meals, and rooms for laborers chasing the next big thing.

Today, though, in