CFJ Op-Ed: Pennsylvania's Data-Center Permission System: When GRID Becomes Command
- Jeffrey Depp
- 1 day ago
- 3 min read
When GRID Becomes Command: Pennsylvania’s Data-Center Permission System
Committee for Justice Senior Fellow Jeffrey Depp has published a new essay in The National Law Review, “Pennsylvania’s Data-Center Permission System: When GRID Becomes Command.”
The essay is the latest installment in Depp’s continuing application of Austrian economics and Public Choice theory to artificial intelligence, energy, and industrial policy. It examines how Pennsylvania has transformed its Governor’s Responsible Infrastructure Development Standards—better known as GRID—from a purportedly voluntary incentive program into an expansive system of administrative permission.
From Incentives to Permission
When Pennsylvania first announced the GRID Standards, officials presented them as the price developers would pay to receive tax incentives, expedited permitting, and coordinated state support. A developer could reject the bargain and proceed under generally applicable law.
That distinction has now largely collapsed.
Gov. Josh Shapiro’s Executive Order 2026-05 makes GRID the organizing principle of Pennsylvania’s data-center permitting system. Developers that accept a project-specific consent order can receive more favorable permit sequencing and retain access to the state’s data-center tax exemption. Those that decline face slower, more rigid review and exclusion from important permitting programs.
A formal choice remains, but the state has loaded the alternatives.
As Depp explains, Pennsylvania has not enacted a conventional moratorium. It has created something subtler: a permission system capable of making the disfavored path slow, uncertain, expensive, and commercially impractical.
The Knowledge Problem in a Consent Order
This essay continues arguments developed in Depp’s earlier National Law Review articles, “Powering AI: Why Markets Matter More Than Mandates” and “Pennsylvania’s GRID Standards and the Fatal Conceit of AI Industrial Policy.”
The central Austrian insight is not that every proposed data center is beneficial or that every developer’s forecast should be accepted. It is that knowledge about the right project, technology, location, financing structure, and power arrangement is dispersed—and often does not exist until entrepreneurs test competing possibilities.
Pennsylvania’s permission system frustrates that process. Data-center development requires coordination among landowners, utilities, generators, equipment manufacturers, investors, customers, and local governments. By forcing essential workstreams into a politically prescribed sequence, the Commonwealth may prevent market participants from discovering whether the pieces can fit.
Meanwhile, scarcity is already inducing adaptation. Businesses are developing more efficient power-and-cooling systems, investors are repricing risk, utilities are revising large-load tariffs, and manufacturers are expanding production of grid equipment. These responses were not designed in Harrisburg. They are emerging from entrepreneurial discovery.
Public Choice and the Politics of Delay
Public Choice theory helps explain why politicians nevertheless prefer visible restrictions.
Local objections to data centers are immediate and concentrated. The costs of a project that never proceeds are dispersed and largely unseen: the power plant that is never financed, the supplier that never expands, the apprentice who is never hired, and the innovation that never occurs.
Politicians receive an immediate reward for announcing that they have stopped an unpopular project. The economic costs arrive later and are borne by people who may never know what was lost.
The temptation is national and bipartisan. New York calls its restriction a moratorium. Texas calls its intervention an audit. Pennsylvania calls its system a consent order. The legal mechanisms differ, but each gives political officials greater discretionary control over private investment without producing the generation, transmission, or infrastructure that scarcity demands.
General Rules, Not Administrative Favor
Legitimate concerns about data centers do not require industrial planning.
Utilities can require large customers to substantiate demand, post deposits, meet development milestones, finance dedicated infrastructure, and bear the costs of cancellation. Environmental and land-use rules can establish measurable limits for noise, emissions, water, traffic, and setbacks. Those rules should be general, prospective, and applied consistently to comparable facilities.
Government should protect property rights, enforce contracts, prevent cost-shifting, and police measurable harms. It should not prescribe the generation mix, wage structure, investment scale, or preferred contractual arrangements.
Once neutral rules are established, prices, contracts, investment, and entrepreneurship should determine which projects work.
The Bottom Line
Pennsylvania once had the right diagnosis: artificial intelligence will require more power, infrastructure, and investment. Its mistake was first to prescribe the path—and then to condition government permission on following it.
A moratorium builds no generation. An audit repairs no transmission line. A consent order shortens no interconnection queue. Each controls the applicant without solving the scarcity.
Pennsylvania can be a laboratory for entrepreneurial discovery or a laboratory for administrative control. It cannot be both.
Read “Pennsylvania’s Data-Center Permission System: When GRID Becomes Command” in The National Law Review.
