The right used to mean “local control” — the principle that the people who have to live with a decision should make it, against the centralizers in Washington or on Wall Street. Then the same right discovered a constituency of billionaires who wanted to build hyperscale data centers in places with weak bargaining positions, and suddenly “local control” meant letting Amazon write the deal with the county board of three.
John Arnold, co-founder of Arnold Ventures, chairman of Grid United, and a sitting member of Meta’s board, published “From New York to Texas, states should respect local control on data centers” on Fox News Opinion arguing that state-level moratoriums on data centers are misguided, that local communities should have the right to decide whether hyperscaler projects are worth the disruption, and that the power-grid fears driving the backlash are overblown. He cites a Lawrence Livermore National Laboratory study finding no connection between data-center growth and retail electricity prices, notes that utilities in nineteen states already impose large-load tariffs, and points to a White House deal under which Amazon, Google, Meta, and more than two hundred other firms agreed to pay for the power infrastructure their facilities require. He celebrates deals: Lancaster, Pennsylvania secured noise limits and twenty million in community-development commitments; Cedar Rapids, Iowa got wage guarantees; teachers in Richland Parish, Louisiana received bonus checks of up to fifty-one thousand dollars from higher sales-tax revenue. State-level bans, he writes, carry “a whiff of luxury belief.”
He is half right. He is half wrong in a way that is the entire story.
Arnold is right that a state moratorium imposed from Albany or Austin on a willing community is the wrong instrument. Blanket bans imposed on communities that want the projects are a mistake. To his credit, the hyperscalers have in many cases agreed to pay for the power infrastructure their facilities require. The tax revenue lands in places that have nothing else coming, and the examples he names are real. I will grant him every one of those points. The rebuttal is to what he left out.
What he left out is what the right used to call local control.
I used to trade agricultural futures on a Chicago desk. I knew the same dynamic from the trading desk. In commodities, a small number of very large firms negotiate one-on-one with a much larger number of small counterparties who each have to take the deal or lose the bid. The small counterparty never has the time, the lawyers, or the data the large counterparty has. The “negotiation” is a contract drafted by the other side’s lawyers, presented as a take-it-or-leave-it. The farmers I grew up around knew that. So did the co-op managers. So did the rural electric cooperatives that built the grid in this county in the first place, because investor-owned utilities said we were not profitable enough to serve.
That is the principle Arnold invoked and walked past.
The Coase Theorem — the intellectual load-bearing wall underneath his argument — assumes parties can bargain to efficient outcomes when transaction costs are low and property rights are clear. There is no transaction-cost theory on earth under which a five-thousand-person county with a part-time lawyer and a three-member board bargains at parity with the firm that controls Meta’s compute infrastructure. The hyperscaler has a permanent Washington lobbying operation, an in-house legal department the size of the county’s workforce, and a press strategy ready the moment anyone objects. Arnold’s version of local control is a vending machine: the range of options is pre-set by the people who built the machine. He offers no account of what makes a “negotiation” real when one party is a trillion-dollar corporation with a team of energy lawyers and the other is a county board whose members hold day jobs and whose attorney moonlights in family practice.
The canal-and-railroad analogy he reaches for carries partial weight — towns that resisted the nineteenth-century rail paid for a generation, bypassed by the commerce that built the country. The analogy breaks at a deeper level. The canals and the railroads were public infrastructure built under eminent domain, regulated as public utilities, with rates and routes set by public commissions and rights-of-way crossing public land. A hyperscaler is none of these things. It is a private corporation under private contracts, selling compute to remote buyers, capturing the upside of public-grid subsidies, and leaving the substation, the water table, and the rate base behind when its compute cycle turns. The canal era built a public inheritance the country still runs on. The hyperscaler builds a private one and asks the county to host it.
And what of the terms Arnold celebrates? Lancaster’s twenty million sounds impressive until you ask what the data center’s total capital expenditure is and what share of that the community actually captured. Richland Parish teachers got fifty-one-thousand-dollar bonus checks; the question is whether that sales-tax revenue will survive the facility’s eventual tax abatement — the incentive packages towns in their desperation routinely offer to attract exactly the projects Arnold says they should welcome. The examples prove that negotiation is possible. They do not prove the community got its fair share, or that it will retain that share across the facility’s thirty-year life, or that it had the institutional capacity to understand what it was signing. Arnold does not ask these questions, and the omission is the argument.
The deepest tell is the word “luxury.” Arnold accuses state-level restrictionists of imposing their preferences on communities that want the jobs. A man with a seat on Meta’s board does not have standing to call skeptics of a boom his own board helped underwrite luxurious. From inside the engine room, the heat is just the heat, and the people who live nearest to it are not the people who set the thermostat. The preferences being imposed are not those of abstract environmentalists but those of a concentrated technology sector that needs rural land, rural power, and rural acquiescence, and that has the resources to get all three.
What is missing from the entire piece is any account of the institutional infrastructure that would make genuine local control possible. A community is more than a counterparty in a transaction; the people in it are owed standing before the agreement is signed, not afterwards as a deduction in someone else’s quarterly return. A town facing a hyperscaler needs independent technical counsel — power-systems expertise that does not come from the company proposing the project. It needs legal representation that can read a community-betterment agreement against the firm’s own fine print. It needs a planning commission that understands zoning not as a bureaucratic formality but as the skeleton of a community’s self-governance. It needs, above all, a countervailing institution — a cooperative, a mutual, a publicly accountable utility — that gives the town standing equal to the corporation it is negotiating with. Arnold assumes none of this exists and proposes nothing to build it. Local control without local power is a courtesy the hyperscaler extends and revokes at will. The community that signs a betterment agreement today may find its grid capacity auctioned to a second facility tomorrow, with no recourse and no renegotiation — because the agreement did not anticipate the next demand, and because the town did not have the institutional muscle to insist on a provision that would.
I want to be specific about what this costs, because the cost lands where the cost always lands.
A hyperscale data center consumes as much electricity as a mid-sized city. The power has to come from somewhere. In rural Wisconsin — and in a lot of rural America — the power comes from a member-owned rural electric cooperative. Adams-Columbia Electric Cooperative, headquartered two miles from where I write this, has more than thirty thousand members across twelve counties. Its members built that grid with their own capital, under the Rural Electrification Act, because the investor-owned utilities said they were not worth serving. They paid for it. They own it. And under the deal Arnold is defending, they will be asked to sell the surplus to a hyperscaler at terms the hyperscaler writes — and to do so in the name of “local control.”
That is not local control. That is local capitulation dressed in the language of subsidiarity.
The principle the right used to defend — that the small can stand against the large, that the place can stand against the abstraction, that the community can refuse what the spreadsheet demands — has been quietly redefined as: the community may not be helped by the state in saying no to the spreadsheet. That is the inversion. The party of Edmund Burke has become the party of the Vanderbilt who bought the legislature.
The model is the answer here, too.
A community does not have to choose between hosting a data center and refusing one. It can own one — not as a tax recipient, but as a member-owned partner in the compute, the way Adams-Columbia’s members already own the poles and the lines and the substation. Adams-Columbia, or any of the rural cooperatives that built the grid under the REA, could be the counterparty in that deal — owning the generation, owning the compute, contracting with the hyperscaler on the co-op’s terms, not the hyperscaler’s. Distributed compute on the rural cooperative grid. Member ownership of the compute as well as the power. The hyperscaler becomes a tenant, not a principal. The community captures the surplus, not just the sales-tax receipts.
That is local control. That is what the right used to mean. That is what the right has, in this case, sold.