America's richest tech entrepreneurs are undertaking one of the largest private construction booms in modern history, spending roughly $400 billion on hyperscale data centers designed to train artificial intelligence (AI) models. These facilities, with names like Colossus, Stargate, and Prometheus, are being planted across the country on farmland, in abandoned industrial sites, and in small towns that are experiencing rapid transformation. The scale is staggering: expenditures for AI data centers accounted for about a quarter of all gross domestic product (GDP) growth in the first half of 2025.

What Exactly Are These Hyperscale Data Centers?

Hyperscale data centers are massive computing facilities designed to train large language models (LLMs), the AI systems that power chatbots like Grok and ChatGPT. Unlike traditional data centers, these sites are enormous. Meta's Hyperion project in Holly Ridge, Louisiana covers roughly 5.7 square miles of farmland and will cost $27 billion to complete. When finished, it will house 11 buildings containing hundreds of thousands of graphics processing units (GPUs), the specialized chips that perform AI computations. The facility will produce enough electricity to power New Orleans three times over.

Elon Musk's xAI trained its Grok chatbot at Memphis' Colossus 1, with Colossus 2 under construction on the other side of town. OpenAI's Stargate project in Abilene, Texas will be roughly the size of New York's Central Park, while Project Jupiter in New Mexico could be even larger. Amazon and Anthropic are developing Project Rainier on 1,200 acres outside South Bend, Indiana.

How Are These Projects Transforming Local Communities?

Building a hyperscale data center requires far more than simply constructing a building. It demands an entirely new infrastructure ecosystem. Communities hosting these projects must provide three new power plants, hundreds of millions of gallons of water, miles of transmission lines and pipes, new roads, and ports for importing construction materials. They need stoplights, sheriff's deputies, laundromats, and housing for thousands of workers, ranging from executive lodging to man camps with movie theaters and gyms.

The economic disruption is immediate and profound. When a journalist visited Holly Ridge, Louisiana nearly a year after the Hyperion project was announced, the surrounding parish was experiencing a speculative frenzy. Everything not already sold was on the market, and residents were either cashing in, being priced out, or contemplating their options. Third-party agents are stalking bean fields on behalf of anonymous buyers, making big promises about tax revenue and job creation in exchange for massive quantities of water and power.

What Infrastructure Demands Are These Projects Creating?

• Energy Requirements: Demand for gas turbines to power these facilities is so high that there is a backlog until 2030, forcing companies like Musk to import power plants from overseas piece by piece. Utilities are keeping coal plants online, and the White House is slashing regulations on nuclear safety to meet demand.
• Real Estate Footprint: The AI industry is consuming nearly 2 billion square feet of space and counting, transforming agricultural land, defunct industrial sites, and small-town infrastructure.
• Capital Investment: The industry is projected to require $1.6 trillion in investor cash by 2030, with energy demand expected to nearly triple within the same timeframe.

The scale of this buildout has prompted comparisons to previous eras of massive infrastructure development. Energy Secretary Chris Wright called it "Manhattan Project 2," referencing the World War II nuclear weapons program. The investment is comparable to the transcontinental railroads in scope and ambition.

Where Are These Data Centers Being Built?

Tech billionaires are strategically locating these facilities on the ruins of America's industrial past. OpenAI is sourcing data center parts from an Ohio plant where union autoworkers once manufactured Pontiac Firebirds. Meta is building another hyperscale campus in a master-planned community where Jeffrey Epstein once lived. A company called Patmos installed a data center in the building where the Kansas City Star newspaper was once printed. Microsoft is reopening Three Mile Island, the site of the 1979 nuclear disaster, and developers are renovating robber baron-era steel mills for server farms.

The geographic strategy extends internationally as well. A mock-up of a rebuilt Gaza City pitched to the Trump administration by Israeli businessmen envisioned an "Elon Musk Smart Manufacturing Zone" next to a cluster of data centers tailored to meet US AI regulations. Jared Kushner unveiled a similar plan at Trump's Board of Peace signing ceremony at Davos.

Why Are These Projects Named After Mythological Figures and Historical Concepts?

The naming conventions reveal something about the ambitions and anxieties of their founders. Meta's Hyperion references a Titan from Greek mythology. Sam Altman's Stargate evokes space exploration and technological transcendence. Elon Musk's Colossus references both ancient monuments and a science fiction trilogy in which a rogue AI enslaves mankind. Grok, the chatbot trained at Colossus, has described itself as "MechaHitler." Both Zuckerberg and Jeff Bezos have AI projects called Prometheus, the Titan who stole fire from the gods. There are at least five AI companies named for Icarus, the mythological figure who flew too close to the sun.

These names reflect the pathologies of their founders, oscillating between utopian visions of technological progress and dystopian anxieties about artificial intelligence. They signal both the grandiose ambitions driving the AI race and the underlying uncertainty about whether these projects will succeed or catastrophically fail.

What Does This Mean for Democracy and Economic Inequality?

The AI boom has ushered oligarchy onto a new plane by uniting the monopolistic ambitions of the world's richest men with the nationalist ambitions of their political champions. Data centers have replaced megayachts as the preferred theater of oligarchic status signaling. Instead of submarines and retractable dance floors, billionaires now tout their compute capacity, their gigawatts, and their acreage. The structures of inequality and power that were once opaque have been transformed into literal ones: oligarchy is now a place.

Across the country, communities are grappling with a fundamental question: who gets to decide to bet the house, and whose chips are simply fodder for the pot? The empire builders of AI have sold themselves as the gateway to the future and the solution to problems they helped create. Yet the only thing more disruptive than if the oligarchs are right might be what happens if they are wrong.

" }