The Notice in the Paper: The Obscure Law Killing America's Data Centers
The largest data-center project ever proposed in America was killed not on its merits but by a newspaper-advertising statute. How town clerks, referendums, and 530 local laws became the AI buildout's most effective opposition.
By MyAudioBooks.ai ·
Listen free: The Notice in the Paper: The Obscure Law Killing America's Data Centers
On the last day of July, twenty twenty-six, the largest data-center project ever proposed in the United States officially ceased to exist. The Prince William Digital Gateway was to be twenty-three million square feet of computing halls spread across twenty-one hundred acres of Virginia countryside — a gigawatt-scale campus, roughly the peak appetite of a major city's worth of electricity, backed by some of the biggest names in digital infrastructure. It had survived years of planning battles, public hearings packed until two in the morning, and a bitterly divided county commission. What it did not survive was a rule about newspaper advertisements.
To feel the weight of what died, it helps to know where it died. Northern Virginia is the undisputed capital of the data-center world — the region called Data Center Alley, where a confluence of fiber backbone routes, cheap land, and cooperative utilities produced the largest concentration of computing infrastructure on Earth. A substantial share of the planet's internet traffic already passes through Prince William County and its neighbors every second of every day. The Digital Gateway was to be the crown jewel of that alley: not another box on the edge of town, but a dedicated district of halls drawn on a scale that would have reset the industry's sense of what a single project could be. The county's own economic-development machinery had championed it for years. When this project died, it did not die in hostile territory. It died at home.
In March of twenty twenty-six, the Court of Appeals of Virginia affirmed that the rezoning approvals underpinning the entire project were void ab initio — void from the beginning, as if they had never legally happened. The county, the courts found, had violated Virginia Code section fifteen point two dash twenty-two oh four, the statute that requires local governments to advertise public hearings on zoning changes — how many times, in what form, with what content, published in a newspaper of general circulation. The notices Prince William County ran ahead of its December twenty twenty-three vote did not comply with the statute and the county's own advertising rules. No valid notice, no valid hearing. No valid hearing, no valid rezoning. No valid rezoning, no project. The largest data-center deal in America was not defeated on its merits — nobody ever ruled on whether it was a good idea. It was voided because the government skipped a step that exists to guarantee that the public finds out what the government is doing before it does it.
This is not a story about one county's clerical error. It is the story of how the AI buildout — the largest, fastest private infrastructure expansion in American history, the one consuming the electricity this series described last time — is being stopped, delayed, and reshaped by the least powerful people in the legal system: town clerks, planning commissions, and residents armed with procedural law. And it raises a genuinely hard question for everyone, whatever you think of data centers: in a country that increasingly cannot build anything big, who should hold the veto — the experts and the capital, or the notice period and the referendum?
At My Audio Books dot A I, you can create your own audiobooks from prompts, turn your documents into audio, all with one subscription, and store your items in your own personal library.
Start with the statute itself, because its obscurity is the point. Virginia Code section fifteen point two dash twenty-two oh four is the kind of law that governs everything and excites no one: a public-notice requirement, a descendant of the era when the way a citizen learned that the county was about to rezone the farm next door was a classified advertisement in the local paper. The law says that before a local governing body can change zoning — the rules that decide what may be built where — it must advertise the hearing, describing the proposal, the place, the time, in a newspaper of general circulation, a set number of times before the vote. The county's own zoning ordinance adds its own advertising requirements on top. These rules are not red tape in the pejorative sense; they are the legal form of a simple principle — the public gets to know before the decision, not after. Skip them, and the decision was never the public's decision at all.
The December twenty twenty-three hearings themselves had already become local legend for their chaos: meeting rooms packed past capacity, hundreds of residents signed up to speak, sessions that ran deep into the early morning as the supervisors heard testimony on both sides. The substance could hardly have been more contested. Which is exactly what made the notice question so decisive, in retrospect: the county's defenders argued that no one in Virginia could possibly have missed the most publicized rezoning in county history, and the court's answer, in effect, was that the law does not measure fame — it measures compliance. A requirement that exists to protect the uninformed is not satisfied by the fame of the project. It is satisfied by doing what the statute says, the way the statute says to do it.
The Virginia courts applied that principle without flinching. In August twenty twenty-five, Circuit Court Judge Kimberly Irving ruled that the county's notices for the Digital Gateway hearings were defective and the resulting rezonings void. On March thirty-first, twenty twenty-six, a panel of the Court of Appeals led by Judge Stuart Raphael agreed, affirming that strict compliance with the advertising requirements is not a technicality to be excused but a precondition of lawful government action. The developers pursued the case to the state's supreme court — and then, in July, QTS, the lead developer, withdrew its appeal. The project that would have been the largest in the world was dead, killed not by a jury, a regulator, or an environmental review, but by a failure to properly tell the public what was about to happen.
Legal scholars have a name for the entire style of opposition this represents, and it is worth learning: proceduralism — the use of process rules rather than substantive arguments to stop or slow government and corporate action. Proceduralism does not ask "is this data center good?" It asks "was the process followed?" It wins precisely when the substance is politically unwinnable, because process violations are objective, documentable, and cheap to prove. The substance of the Digital Gateway fight was contested — jobs versus landscape, tax base versus transmission lines. The procedure was not contested at all. Either the ad ran correctly or it did not.
Prince William County is the biggest kill, but it is only one rung of a three-rung escalation now underway across the country. The second rung is direct democracy. In April twenty twenty-six, voters in Port Washington, Wisconsin — a small city on Lake Michigan facing its own massive proposed data center — approved what reporting described as the nation's first anti-data-center referendum, by a margin of two to one. The measure does not ban anything directly. It does something more durable: it requires public approval before the city can create large tax-incremental districts, the financing vehicles typically used to subsidize exactly this kind of development. A tax-incremental district, for the uninitiated, is a device that lets a city borrow against the future tax revenue a project is expected to generate, spending money on the project today in anticipation of its taxes tomorrow — which means the public is quietly underwriting the deal years before it sees a dime. Port Washington's referendum takes that device out of the back room: from now on, the public must say yes before its future taxes can be mortgaged for a data center. The residents of Port Washington did not just fight one project; they changed the rules so that the next project cannot be financed behind closed doors.
The third rung is sheer volume. According to tracking by the climate publication Heatmap and confirmed by independent watchdog databases, more than five hundred thirty local laws seeking to ban or restrict data centers were enacted across the United States in twenty twenty-six alone. The bluntest instrument in the wave is the moratorium — a legal freeze on all approvals while a town writes new rules, buying time without committing to any final position. Dozens of communities have declared them, some for six months, some for years, each one a pause button pressed on a billion-dollar timeline.
A second instrument works by distance rather than time: the setback, a zoning rule that requires a structure to sit a fixed distance from homes, schools, or water — hundreds of feet, in the new ordinances, sometimes more than a thousand. A generous setback does not ban a data center outright; it simply prices it out of every parcel that matters, achieving by geometry what a ban achieves by name, with less legal exposure. Add noise ordinances tuned to the very hum this series heard from the access road, and outright bans in the boldest jurisdictions, and the toolkit is complete.
Then there are the fights that turn on the map itself. In Menomonie, Wisconsin, a resident campaign swelled past ten thousand members and stopped a one point six billion dollar project that had been advancing through closed-door meetings — a deal whose mechanics included annexation, the process by which a city absorbs surrounding land into its own jurisdiction, taking control of its zoning, its taxes, and its destiny in a single vote. Annex the farmland, and the county's skepticism no longer matters; the city's friendly council decides. The Menomonie campaign pressured the mayor and council into halting both the annexation and the rezoning, and the project died of geography as much as politics. In Springdale Borough, Pennsylvania, the council approved a project over furious opposition — and the fight simply moved to the county permitting level, where it continues. The pattern is not a few NIMBY towns. It is a national, bottom-up regulatory surge, assembled one ordinance at a time, faster than any federal or state government has moved on the same issue.
At My Audio Books dot A I, you can listen to this story and thousands of others that explore the hidden science and mechanics behind the headlines.
Why is the industry losing this fight? Not because it lacks money, lawyers, or political connections — it has all three in historic abundance. It is losing because it prepared for the wrong war. The technology industry spent five years bracing for a conflict over what AI does — the safety debate, the jobs debate, the existential debate — and built policy shops in Washington and Brussels to fight it. It did not prepare for a conflict over where AI is allowed to physically exist, a fight that takes place in planning-commission meeting rooms in counties the industry's lobbyists have never visited, under rules written generations before the internet. The AI buildout's political defense is centralized, sophisticated, and aimed at capitols. The opposition is distributed, procedural, and aimed at zoning boards. The two armies are not even on the same battlefield.
There is a second, subtler reason: secrecy breeds its own antibodies. Data-center projects are routinely negotiated under nondisclosure agreements, with code names and shell entities, on the theory that quiet deals face less resistance. Menomonie is what happens when that theory fails: the discovery of a secret one point six billion dollar negotiation did not lower resistance, it weaponized it, handing organizers the narrative that something was being hidden from the public — because it was. The industry's instinct for quiet, entirely rational deal by deal, is generating the single most powerful organizing story the opposition has.
Now the strongest case against the backlash, stated as strongly as it deserves, because it is not nothing. The data-center opposition has real costs, and somebody pays them. Prince William County did not merely lose a campus; it lost a tax base measured in the hundreds of millions of dollars, in a state where data centers fund schools and hold down residential property taxes. Landowners who wanted to sell into the project had their property's value legally trapped by their neighbors' preferences — the county's error aside, the underlying fight was about what an owner may do with private land. The procedural win, for all its democratic poetry, required no one to ever answer the substantive question of whether the project should exist, and proceduralism, once normalized, is available to block everything — housing, transmission lines, solar farms, shelters — not just unpopular megaprojects. A country where every project of scale can be voided on a notice defect is a country that will struggle to build the grid the same residents say they want. There is even a name for the national condition the critics describe: the veto point — every doorway in the system through which a single determined group can kill a project, and America has more of them than any country that still builds big things. The strongest critics of the backlash do not defend the secrecy or the scale. They ask whether a veto this easy to trigger can possibly coexist with the fastest infrastructure buildout the economy has ever attempted — and they note that the electricity has to come from somewhere, whether or not the substation is in your view.
The industry, for its part, is not standing still. Its countermoves are already visible, and they follow the logic of the loss. If the battlefield is local, move to friendlier ground: developers are rerouting to counties and states competing to offer pre-approved sites and guaranteed timelines. If the rules are the problem, change the rules: industry groups are lobbying state legislatures for preemption — state laws that strip zoning authority over data centers from local governments entirely, on the argument that regionally significant infrastructure cannot be subject to a municipal veto. Preemption is not a new invention, and its track record is exactly why both sides take it seriously. States have used it to override local objections to pipelines, to renewable energy siting, to fracking, to broadband — each time with the same logic: infrastructure that serves millions cannot be hostage to the nearest thousand voters. Each time, the locals answer with the same logic too: the people who live with the noise and the transmission lines are not an externality to be managed. The data-center fight is about to become the largest test of that question in a generation. And if the grid connection is the vulnerability, bypass it: the newest campuses increasingly arrive with their own on-site power, gas plants and eventually nuclear reactors, reducing their dependence on the local utility politics that opponents exploit. Each of these responses will generate its own opposition, at a different level of government. The fight is not ending. It is moving up a floor.
What would prove this article's account wrong — what would show the backlash is a passing wave rather than a durable restructuring? Three findings would disprove it, stated concretely. First, if state preemption laws begin passing in the data-center-heavy states — if legislatures in Virginia, Georgia, or Texas strip local zoning authority over these projects — the town-clerk era ends as suddenly as it began, and the fight relocates to statehouses where the industry's centralized lobbying is strongest. Second, if the procedural wins start being reversed on further appeal or by legislative cure — special sessions that retroactively validate flawed approvals — the notice-rule weapon loses its edge, and the opposition is forced back onto substance, where it is weaker. Third, if the pace of the local-law wave itself slows — if the count of new moratoriums and restrictions stops compounding — then twenty twenty-six will look like the peak of the backlash rather than its beginning.
It is worth saying what this article has not claimed. It has not claimed the Digital Gateway project should have been built, or that it should not; the courts never decided that either, and neither does this article. It has not claimed Prince William County acted corruptly — the record shows defective notice, not bribery, and the distinction matters. It has not claimed the five hundred thirty local laws are all wise; they range from careful planning to pure obstruction, and the difference between those is the hardest question in the story. And it has not claimed the industry is wrong to want certainty — the demand for power in the previous article of this series is real, and it will be built somewhere. The only question is who gets a say in where.
Which returns to the newspaper advertisement, because there is something almost poetic about the instrument that stopped the largest data-center project in the world. The industry runs on the fastest information networks ever built, moving zettabytes across fiber at light speed. Its opponents stopped it with a rule written for classified ads — the requirement that before the government changes what may be built next to your home, it must tell you, in public, in a form you can find, in time for you to show up and object. That rule is older than the internet, older than the television, older than most of the counties that enforce it. The machines being built in those windowless halls are supposed to be the future. The notice in the paper is the past — and this year, in court after court, meeting after meeting, the past is winning.
The open question is whether that is a restoration or a rot. The answer probably depends on what gets built near you, and who told you first.
At My Audio Books dot A I, you can create fiction, non-fiction, and turn your documents into audio, all stored in one place with a single subscription — plus get instant access to thousands of audiobooks and deep-dive investigations. Learn more today at My Audio Books dot A I.