Thoughts on the Market
2026-08-05 · Hosted by Mike Wilson · Morgan Stanley
Executive Summary
This episode, hosted by Ariana Salvatore with guests Michelle Weaver and Sarah Wolfe, examines the politics and economics of the U.S. AI data center build-out. More than 300 local moratoria on data centers have passed since 2023, touching 40 states, as communities weigh construction-phase job gains against longer-term costs like higher electricity and housing prices.
Key Stories & Changes
1. Local Opposition to Data Centers Is Spreading but Rarely Permanent
More than 300 local moratoria have passed since 2023, now touching 40 states
Most are temporary pauses (often one to three years) for impact studies rather than outright bans
Opposition spans both Democrat- and Republican-held state legislatures, reflecting bipartisan concern over electricity prices and quality of life
2. The Local Economic Trade-Off: Jobs Now, Uncertain Benefits Later
A 250,000-square-foot data center in Virginia supports more than 1,500 construction jobs, but only about 50 full-time positions once operational — and Wolfe called that a high-end estimate, with other states reporting as few as 10-30 permanent jobs
County-level research shows positive effects on employment, business formation, wages, income, and tax revenue during construction, but the labor multiplier fades sharply after the facility comes online
Durability of tax revenue benefits depends on how localities structure incentive packages; overly generous, long-lived exemptions can erode the fiscal upside that made a project attractive
3. Power Bottleneck: A Projected 38 Gigawatt Shortfall
Morgan Stanley estimates a potential shortfall of around 38 gigawatts of power needed through 2028
A typical AI data center can use as much electricity as 100,000 homes
The power bottleneck is pushing hyperscalers toward off-grid, behind-the-meter power generation to avoid visibly impacting consumer utility bills
Skilled labor shortages, especially electricians, are compounding the buildout bottleneck
4. Geopolitics Makes a Federal Ban Unlikely
China is reportedly spending roughly $300 billion over five years on its own domestic AI data center build-out
Researchers have identified China-linked influence operations using American frontier AI models to generate content linking data center construction to rising energy prices
Federal lawmakers, unlike state and local officials, weigh AI infrastructure development against U.S.-China strategic competition, making a nationwide ban or moratorium "very unlikely"
Policy tools currently favoring continued build-out include permitting reform and use of the Defense Production Act
5. Geographic Shift Toward Rural Sites and International Locations
Rising local resistance is shifting the geography of new data center construction toward rural locations
More capacity is being built in Canada and Australia specifically to serve U.S. AI demand, reflecting a form of "nearshoring" for compute capacity
Trends Identified
1. From National Debate to "Conditional Build-Out"
Rather than expecting a uniform national resolution, Morgan Stanley's strategists see the AI infrastructure build-out settling into a project-by-project negotiation model, where hyperscalers offer specific concessions — like on-site power generation or improved utilization rates (currently just 30-40% at some facilities) — to secure local buy-in. This decentralized dynamic will likely slow the build-out's pace without stopping it.
2. Perception, Not Just Data, Is Driving Political Backlash
Michelle Weaver noted that consumer perception that data centers are raising their power bills is spreading even into regions with regulated utility markets where the linkage is weaker. This gap between actual cost pass-through and public perception is becoming a significant political variable, independent of the underlying economics.
3. National Security Considerations Are Overriding Local Economic Concerns at the Federal Level
The strategists framed U.S.-China AI competition as a structural reason federal policymakers will continue prioritizing build-out speed over local objections — creating a persistent tension between state/local resistance and federal facilitation that is likely to persist through and beyond the midterm elections. ---
Sentiment Analysis
Overall Market Sentiment: Cautiously Constructive
The tone is analytical rather than bullish or bearish on markets directly, but the strategists expect continued elevated AI capex (money spent building data centers and infrastructure) despite rising political friction.
Risk Factors Highlighted
Power shortfall: A projected 38-gigawatt shortage through 2028 could constrain the pace of the AI build-out.
Local moratoria proliferation: Over 300 moratoria across 40 states could delay or reroute projects, even if most are temporary.
Consumer electricity cost backlash: Rising perception (accurate in some regions, less so in others) that data centers raise utility bills is fueling political resistance.
Incentive package design risk: Overly generous local tax exemptions could erode the fiscal benefits that justified a project in the first place.
Skilled labor shortages: A shortage of electricians and other skilled workers is limiting the pace of construction.
Geopolitical influence operations: China-linked campaigns are reportedly using AI models to amplify anti-data-center sentiment tied to energy costs.
Front-loaded job benefits: Permanent employment after construction (10-50 jobs per facility) is far smaller than the construction-phase boost, risking local disappointment over time.
This episode was covered in today's [The Market Signal — 2026-08-05](https://marketsignal.beehiiv.com/p/the-market-signal-2026-08-05), a cross-source synthesis of multiple podcast reports.