AI Data Centers Are America’s Key to the Future—Not the Bogeyman

Artificial intelligence is a generation-defining technology with applications spanning defense, health care, law, and finance. In the coming decade, it will determine which companies dominate markets and which nations shape the global order.

With China making state-backed investments at scale, the US can’t afford to fall behind. Preserving leadership requires accepting a simple reality: AI depends on physical infrastructure. Sustaining it means massive new investments in data centers.

Data centers are large warehouses of servers that power the computations behind large language models, which are the backbone of AI systems such as ChatGPT, Gemini, and Claude. The US now hosts more than 5,388 data centers, nearly double what it hosted in 2021. According to the Department of Energy, they consumed 4.4% of the nation’s electricity in 2023, a share that could climb to as high as 12% by 2028.

The reality is unavoidable: AI depends on data centers, and data centers depend on unprecedented levels of energy production.

Because data centers must have a physical location, they’re a target for opposition to AI and its energy demands. Local backlash has derailed projects before they even begin, with residents lobbying city councils to block rezoning efforts, deny tax abatements, or withhold construction permits. What might appear as routine land-use disputes have become chokepoints for the future of US AI competitiveness.

In Missouri, for example, data center projects in St. Charles and Peculiar were derailed after public backlash, sinking multibillion-dollar data center ambitions. St. Charles went even further, enacting a one-year moratorium on new data centers—the first such ban in the nation.

In Memphis, Tenn., Elon Musk’s xAI data center has sparked community opposition after it was alleged that unpermitted methane gas turbines contributed to higher carbon dioxide and nitrogen dioxide levels, worsening pollution and respiratory problems in nearby communities. The City of Memphis conducted its own review and determined the air was safe.

These fights follow a familiar script: Critics argue data centers provide few permanent jobs, put a strain on the energy supply, drive up utility bills, and in some cases, cause noise pollution.

The reality is far more complex.

At the front end, data center projects deliver massive investment to communities. Meta has invested $1.5 billion in data centers across Tennessee, creating thousands of construction jobs and hundreds of permanent jobs. Meta also invested $1.4 million in local schools while building its Gallatin, Tenn., campus, and it claims its “data centers’ electricity use is matched with 100% clean and renewable energy.”

In Memphis, the xAI data center campus transformed a vacant power plant into high-value infrastructure and is expected to contribute $25 million annually in new property tax revenue to the city, which has struggled with one of the nation’s largest population exoduses in recent years.

Across the country, these facilities have been paired with investments in education and workforce training. In Kansas City, Mo., Meta included investments in local schools and STEM programs in its $1 billion data center project. Even with property tax abatements, communities benefit from increased sales and utility taxes, along with the broader revitalization that comes when dormant land becomes a hub of critical infrastructure and attracts new businesses and job creation.

Another critical nuance in this debate is that not all data centers are the same. Earlier this year, a 12-megawatt, 102,500-square-foot “colocation” data center broke ground in Nashville. Designed to host multiple tenants, it will provide secure cloud capacity for technology businesses.

By contrast, across the state in Memphis, Elon Musk’s xAI project is a “hyperscale” data center with a reported 300-megawatt capacity and nearly 785,000 square feet—home to what is touted as the world’s largest supercomputer powering the Grok chatbot.

This isn’t to say concerns are misplaced. Hyperscale AI campuses can draw hundreds of megawatts, requiring significant grid upgrades. Cooling demands raise questions in water-stressed regions, and poorly structured utility contracts can shift infrastructure and utility costs to households and small businesses.

These are real issues, but they are regulatory challenges—not reasons to reject data centers outright. Some states, for instance, are already taking steps to create separate utility rate classes for data centers to ensure large-load customers pay their fair share rather than leaving ratepayers to shoulder the burden.

The broader picture can’t be ignored. Data centers aren’t optional luxuries; they are the nervous system of the digital economy and the backbone of AI training and development.

The real question isn’t whether to build them, but where and under what conditions. Communities should insist on transparency in tax arrangements, regulators should ensure utility fairness, and companies must continue investing in efficiency and renewable energy. Wholesale opposition is short-sighted.

AI will define the 21st century. If the US treats data centers as the bogeyman, local fights over zoning and utility bills will become a national liability, opening the door for competitors to surge ahead.

Data centers must go somewhere. Building them responsibly here is far better than letting AI’s physical infrastructure migrate elsewhere. The stakes aren’t simply about electricity and land use. They’re about whether the US remains at the forefront of the next technological era or cedes that position to others.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.

Author Information

Oliver Roberts is an adjunct professor of law at Washington University in St. Louis School of Law, co-head of Holtzman Vogel’s AI practice group, and founder and CEO of Wickard.ai.

Write for Us: Author Guidelines