My family moved to Loudoun County, Virginia, in 2005. Dotted with farms, horses and wineries, the region was a growing part of Northern Virginia’s IT industry. Two decades later, Loudoun is one of the wealthiest counties in America and widely known as the “data center capital” of America. It is also a part of the national backlash against the data center industry. In Leesburg, the announcement of a new data center close to my parents’ community spurred some residents to hang “for sale” signs in their front yards. HOA meetings and public hearings were full of residents worried that construction would lower property values and that a giant commercial building would obscure the view.
The backlash is happening nationwide: 71% of Americans now oppose having a data center in their area. Neighboring Prince William County rejected a 2,000-acre development proposal. Politicians from Pennsylvania to Texas are pivoting with the political winds. Many who once championed data center investment have turned their backs on a political liability in an election year.
Mere NIMBYism cannot explain this anger. More than anyone else, Silicon Valley created the conditions for the backlash. Earlier waves of technology sold people things they wanted. Amazon promised cheaper books and convenience. Facebook and YouTube promised connection and an audience. Much of AI’s sales pitch is aimed elsewhere. It promises employers efficiency — meaning fewer people on the payroll — while telling everyone else the change is inevitable and they should simply adjust.
Since ChatGPT’s release in 2022, AI executives have framed their technology as miraculous and catastrophic. Anthropic CEO Dario Amodei described one possible future: “Cancer is cured, the economy grows at 10% a year, the budget is balanced, and 20% of people don’t have jobs.” Whatever he intended, the public heard a CEO predicting mass unemployment while his company raked in profits. Most Americans encounter AI through headlines about layoffs, surveillance, and systems beyond human control — not as a helpful coding tool or a medical breakthrough. That is the transformative vision AI leaders have promoted: billions for Big Tech, universal basic income for those laid off.
Silicon Valley lacks the self-awareness to see how the rest of America views it. The hackers who once disrupted traditional companies with code and algorithms are no longer rebels. The counter-elite are now the establishment, with enormous wealth, political influence, and direct connections to the White House and the Pentagon. For many Americans, data centers are the physical presence of the tech monolith in their lives. Their anxieties about unemployment, economic disruption, and social control can be projected onto a windowless building down the road.
Data center developers had long relied on “speed, confidentiality and political connections” to secure approvals on favorable terms. They sought subsidies behind closed doors and asked local officials to sign NDAs. Some confidentiality is defensible. Google does not want Meta to know where its next facility will be built, and a farmer may add another zero to the asking price after learning who wants the land. NDAs also kept towns from comparing offers and residents from organizing against a secret corporation that held all the cards. To the public, it looked like the fix was in.
The industry compounded its secrecy with bad political judgment. Former senator Kyrsten Sinema’s campaign for a data center in Chandler, Arizona illustrates the problem. A former politician with little credibility among MAGA voters or progressives came into town to sell the AI revolution. Her method of persuasion was warning locals that Washington might force the project on them anyway. When it was time to vote, residents packed City Hall, and the Chandler City Council rejected it 7-0.
If the industry wants anyone to listen again, it has to alter its strategy. Companies worth hundreds of billions cannot plausibly cast themselves as victims of irrational locals. The industry should stop the apocalyptic unemployment predictions, stop operating in secret, and start dealing with communities as partners instead of obstacles.
The AI economy isn’t made up only of software engineers, AI models, and venture capital. A better strategy should promote the people who are building the physical parts of the AI economy. That is a far better story than predicting radical societal change. In a New York Times op-ed, union official Don Slaiman of IBEW Local 26 wrote that his members went from 14 million hours of work a decade ago to 28 million in 2025. The local added 600 apprentices last year. The Wall Street Journal has reported similar opportunities for welders, plumbers, and electricians earning more than $100,000 during the data center construction boom.
The construction boom will not last forever. But these campuses take years to build, and the work continues as they are expanded, rewired, cooled, and maintained. Each project builds up a talent pool of tradespeople capable of handling increasingly complex electrical and industrial systems. That capacity is as much a part of America’s technological strength as its microchips and frontier models.
The people telling the story matter. An apprentice can talk about steady work for his career. A superintendent can point to school renovations that tax revenue paid for. Send those people, not another set of lobbyists who disappear once the vote is over. The wrong approach makes residents think they’re being processed by the Silicon Valley machine. Send people who will still be there when the cranes leave.
Every project should come with a fair local deal, with large users paying their full share of grid costs and the nearest homeowners receiving noise mitigation and utility credits. Subsidies should flow only after local hiring targets and water limits have been met and enforced. Governors also cannot spend years courting data centers and then blame the industry for revealing how little generation and transmission capacity their states have built.
Only after the local bargain is clear does the national-security argument deserve a hearing. We cannot compete with China in AI while treating AI infrastructure as illegitimate. Power plants, transmission lines, and substations are what keep the lights on. Fiber, transformers, and data centers are what keep the internet running. Trying to ban all data centers will push the capacity overseas, along with Americans’ data. Local governments should be able to reject bad proposals. They should not be able to stall every project indefinitely once the developer has met clear, public rules.
The backlash after Three Mile Island helped make nuclear power politically toxic for a generation. The plastic straw panic gave us soggy paper straws while doing little to address plastic waste. Data centers are becoming a convenient target for political failures in economic and infrastructure policy that predate AI. The right answer is a harder bargain with the industry — followed by building the capacity the country needs.
***
Anang Mittal is a Washington, D.C. based public affairs professional. He was a staffer for Speaker Mike Johnson and former Senate Majority Leader Mitch McConnell.
.png)
.png)
