Gen Z’s religious turn, the data center backlash and a ‘lack of consent’: how the youngest adults are winning battles the millennials lost

Opposition to AI data centers has blocked or delayed more than $228 billion of projects since the start of 2025, including $68 billion across 45 projects in April-June and $130 billion in the first quarter. More than 500 localities in 39 states have enacted bans, moratoriums or restrictions, while New York, Pennsylvania, Texas and numerous California jurisdictions have moved to constrain development. The backlash creates material permitting, construction-timeline and capital-deployment risks for the trillion-dollar AI infrastructure buildout, with developers already pursuing litigation against local bans.

Analysis

The investable implication is a rising permitting-duration and power-interconnection premium, rather than an immediate collapse in AI capex. Hyperscalers can redirect projects across jurisdictions, but each relocation extends deployment timelines, raises land/grid costs, and makes previously entitled sites more valuable. This favors existing powered capacity at EQIX and DLR and, selectively, utilities with surplus generation or faster regulatory pathways; it is negative at the margin for greenfield-dependent developers and for near-term data-center equipment order conversion at VRT, ETN and GEV.

The market is likely underpricing geographic concentration risk: a politically hostile county can delay an entire campus, while equipment suppliers retain their revenue only if aggregate hyperscaler capital budgets—not announced projects—remain intact. Over the next 1-3 months, local moratoria, environmental-permit rulings and utility interconnection queues can produce headline-driven volatility in AI-infrastructure suppliers. Over 6-18 months, constraints should shift demand toward brownfield industrial sites, behind-the-meter generation, storage and states willing to trade permitting certainty for tax revenue; PWR is relatively insulated because grid-hardening and transmission spend persists even if individual campuses move.

Contrarianly, opposition may be bullish for incumbent data-center REIT economics and for power assets, not broadly bearish for AI. Scarcity supports pricing and can force hyperscalers toward leased capacity or premium power contracts. The thesis fails if hyperscalers materially cut total capex rather than reroute it, or if state/federal preemption streamlines siting; monitor VRT/ETN backlog conversion, EQIX/DLR leasing spreads, utility load forecasts and announced capex from MSFT, AMZN, GOOGL and META.

Key Decisions for Investors

• Initiate a 6-12 month pair: long EQIX / short VRT in equal dollar amounts. Regulatory delay increases the value of operational, powered capacity while creating downside risk to the timing of VRT project revenue; reassess if EQIX leasing spreads soften or VRT reports backlog cancellation rather than merely delayed delivery.
• Maintain overweight PWR versus data-center-specific electrical-equipment exposure (VRT, ETN) for the next 1-3 quarters. PWR captures transmission, substation and interconnection remediation regardless of the ultimate campus location; reduce if utility capital plans show load-forecast cuts or transmission approvals weaken.
• Use any broad AI-infrastructure selloff to accumulate DLR on a 6-18 month horizon, but only if management sustains leasing and development yields. The scarcity thesis is invalidated by falling occupancy, a meaningful decline in cash leasing spreads, or evidence that hyperscalers are abandoning third-party capacity.
• Set an alert around upcoming state permitting or preemption actions in major data-center markets. A coordinated pro-development response would be a catalyst to cover the EQIX/VRT pair and re-enter VRT/ETN, because equipment names would recover faster once delayed projects regain a defined construction schedule.