The Data Center Authorization Price — A 50-State Baseline

What Every State Charges for AI Data Center Authorization, Where the Price Moves Next, and Why Certainty Outsells Permission

Every AI data center project in America now pays an authorization price — the accumulated legal and operational cost of permission to site, interconnect, power, and energize — and no two states charge it the same way. Twenty-four states hold approved large-load tariffs. Nine carry dedicated enacted authorization floors. More than one hundred localities enforce binding pauses. The rest charge through silence, and silence is the most expensive price of all, because its terms get written after capital commits.

Installment I of the AI Infrastructure Authorization Series prices all fifty states on one instrument, against a July 2026 baseline, and stakes thirty-four dated predictions on the result. [Fallback if the T₀ register is still the live version: "…and anchors a public prediction register that has grown to thirty-four dated entries."] Readers who continue get three things no other document provides: the map of what every jurisdiction actually charges, the mechanism that indicates where each price moves next, and a working playbook for operating across all fifty regimes at once.

Read the full publication.

Two standalone resources work alongside it: the 50-State Authorization Price Atlas — fifty-state tables, per-state detail, the contested-state playbook, the moratorium tracker with expiration calendar, and every prediction contract — and the Model Data Center Authorization Code, the grading instrument that scores any data center law — statute, tariff, executive order, ordinance, or RTO rule — on one strictness spectrum, so every jurisdiction's terms become directly comparable.

Strictness Is the Wrong Variable to Shop On

Most readers arrive asking which states are strict, and the map answers. The paper's core finding shows why the question misleads: the strictest states keep winning the most bidders.

Virginia demands $1.5 million per megawatt in collateral and holds the deepest data center market on earth. Texas mandates curtailment and drew 474 gigawatts of connection requests. The evidence supports the reverse causal direction: capacity attracts load, accumulated load then produces stricter terms within one to two legislative sessions, and among legal variables, certainty appears more commercially decisive than nominal leniency.

The pattern across the full map reinforces the read. Capacity-poor lenient states attract essentially nothing, while hyperscalers maintaining multi-state portfolios still accept Virginia's published price rather than rely on nominally lenient but unwritten regimes. A capital committee can model a written price in an afternoon; no model prices an unwritten county revolt.

A National Term-Sheet Architecture Now Exists

Twenty-four state tariff regimes now share a recurring architecture — large-load coverage classes commonly set between 20 and 75 MW, minimum-take obligations clustering at 85–90%, contract terms running 10–15 years, cost-causation language — arrived at without coordination. Each approved instrument lowers the next jurisdiction's drafting cost: commission staff cite prior dockets, legislative counsel copy enacted structures, utilities file terms their peers already litigated.

Multi-state operators should treat the convergence as the most valuable fact in the dataset. A firm can now begin from one national term-sheet architecture and vary structures and parameters by jurisdiction, rather than negotiating fifty regimes from scratch. Evidence that the terms screen demand is direct: AEP Ohio reported more than 30 GW of pre-tariff interconnection requests narrowing to 13 GW of paid engineering studies and 5.6 GW of signed, collateral-backed contracts once financial commitments attached.

The Model Data Center Authorization Code supplies the measuring stick: every score in the baseline grades against its twelve provision families and published strictness formula.

The Price Has Moved One Way — So Lock Terms Now

Between the 2024 sessions and the July 2026 baseline, no state loosened an enacted term, and the year's loudest restriction defeats removed proposed pauses without touching a single enacted one. Four days after the baseline closed, the pattern escalated: Governor Abbott's audit directive paused the entire ERCOT interconnection queue pending verification — the second-largest data center state executing the strictest single instrument on the map.

Two operational consequences follow, and both reward speed. If the ratchet continues — as the register predicts at 85–92% — today's terms are likely the cheapest available vintage in any given state. And queue position behaves economically like property: a firm's place in Dominion's auto-enrollment, Texas's rule stack, or New York's re-entry framework appreciates with every tightening enacted behind it. Grandfather clauses and tariff vintages are starting guns — the firm that locks terms before each deadline holds an asset its later-arriving rivals must buy at the new, higher price.

How Hyperscalers and Developers Use It Across Jurisdictions

Cross-jurisdiction operation is where the baseline earns its keep, because the market's real structure is fifty depleting regimes raced simultaneously, not fifty stable ones compared once. The paper equips the race five ways.

Price all four layers, not one. A state row is a floor, never the bill. Every site carries a state authorization price, a local overlay, a regional service overlay — the layer through which Monitoring Analytics attributes $6.3 billion of PJM's latest capacity auction, and $29.4 billion across the last four, to data center demand — and a thin federal layer. The baseline scores each separately so a portfolio model can too.

Read the four outputs per site. Strictness, price certainty, authorization friction, and overlays are scored independently for every state, because each answers a different question: what the terms cost, whether they can be known, how hard approval runs, and what sits on top. Virginia and Indiana carry similar headline costs; the four-output read shows Indiana's price is not lower, only later — and later costs more, because repricing after capital commits hands the surplus to whoever writes the late term.

Sequence entries against the expiration calendar. Local moratoria run six to twenty-four months with defined ends, and the tracker maps every computable expiration. An operator arriving at expiration with terms in hand converts every pause into a scheduled negotiation instead of an ambush.

Know which race each tier runs. Bargain-enacted states reward racing for permits and queue slots under current-vintage terms. Bargain-forming states reward racing for the drafting seat — whoever negotiates the template sets rivals' costs for a decade, and the seats are open now in named dockets. Accelerate states conceal a trap: friendliness there is untested and depletes on contact, so the first mover either pays the terms-writing toll amid backlash or captures template authorship. Terms-first firms capture it; announce-first firms pay it.

Operate contested states by rule, not instinct. Sequence county-first. Deliver the enforceable loss-preventing instrument before opposition organizes. Buy the drafting seat at county scale. Run a multi-county portfolio. Never announce without a signed community instrument, and never sue the county. The organizing principle inverts intuition: the county with an adopted ordinance is safer than the county with no rules, because moratoria cluster where local law is silent.

Together the five moves convert regulatory exposure from an unbounded risk into a priced, scheduled variable — which is the entire function of a baseline.

What Every Other Stakeholder Takes From It

Four more seats read the same map, and each leaves with a specific answer.

State lawmakers get the term-sheet architecture two dozen peers already enacted, plus a playbook keyed to position — incumbent magnets defend certainty, power-frontier states enact the bargain before the first campus while leverage peaks, constrained states manufacture headroom and end irresolution, thin markets preserve options instead of pre-fighting arrivals.

Commissions and utilities get the template cascade working in their favor: the next docket runs easier than the last, and terms arriving from peer jurisdictions come pre-litigated.

County councils get the finding the entire local layer turns on — a written ordinance beats silence for both sides, and a moratorium works best as a bargaining instrument with a published conversion path.

Federal readers and investors get, respectively, the layer's honest size — states and counties decide roughly nine in ten approvals, and expedited permitting defers cost rather than cutting it — and the register itself: thirty-four dated claims that settle in public, the tradable version of the analysis.

Thirty-Four Predictions on the Record

The register grades in public through July 2028, with first checkpoints in October 2026. Five headline it: no state that has written terms un-writes them (85–92%) · seven of the next ten major campuses land in published-terms jurisdictions (70–80%) · the Texas audit shrinks the 474 GW queue by nearly a third (78–88%) · enacted authorization floors spread from nine states to fifteen by end-2027 (72–82%) · no federal preemption of any state siting or tariff law (90–95%). Every entry carries a deadline, a falsifier, and a named settlement source, and misses publish at the same prominence as hits.

The Record Behind the Analysis

Every MindCast forecast is timestamped on the public record and validated against outcomes. The standing record includes a Department of Energy comment in the federal docket with three predictions resolving by July 2028; a CFTC filing that requested a rulemaking conversion within ninety days and saw the Commission deliver it in forty-one; and live validation in arenas that settle overnight — Super Bowl LX and the 2026 World Cup Final, modeled and simulated in advance.

Where the Analysis Goes to Work

The baseline is a living, versioned instrument. The 2027 sessions carry a standing watch list, updated versions publish as the record moves, and customized applications extend the same instrument at operational resolution — State Authorization Profiles, multi-state site-selection comparisons, site-specific authorization stacks across all four layers, and governor's competitive-position reports. The public framework and the applied work remain one architecture.

For a State Authorization Profile, multi-state comparison, site-specific authorization stack, or governor's competitive-position report, contact MindCast AI.

MindCast AI produces institutional and behavioral analysis of AI infrastructure governance, built on cognitive digital twins and dynamic predictive game theory. Read the full publication and the standalone 50-State Authorization Price Atlas.

Related MindCast works: Related works: Three Competing Governance Equilibria for AI Infrastructure · Why MindCast Is Filing a Public Comment with the Department of Energy · New York’s Data Center Moratorium · The Two-Ledger Data Center Bargain · The Federal-State AI Infrastructure Collision · The AI Infrastructure Power Stack