These 10 states are best positioned to land AI data center deals despite rising public opposition

CNBC’s 2026 “America’s Top States for Business” infrastructure rankings highlight states best positioned for AI, emphasizing abundant/inexpensive electricity, water, and certified “shovel-ready” data-center sites. Key examples include Ohio ranking No. 1 (A+) for infrastructure and Virginia No. 2 (A+), while Pennsylvania (Amazon’s $20B data-center investment) ranks No. 11 and Arizona’s grid faces water stress alongside $12B/20-year water infrastructure repair needs. The article also flags policy headwinds as Illinois pauses new data-center incentive agreements from July 1 and notes growing local opposition to AI/data centers.

Analysis

The investable read-through is that AI capex is turning into a scarcity trade around power, water, and permitted land rather than a simple “more cloud demand” story. That shifts bargaining power toward firms that can aggregate infrastructure and entitlement risk — especially platform owners like BX and, to a lesser extent, hyperscalers such as AMZN and ORCL that can pre-commit capital to secure scarce capacity. Over a 6-18 month horizon, the value capture should skew to whoever controls the bottlenecks, not necessarily whoever generates the most AI headlines.

Near term, the main risk is that project economics get pushed out by interconnect queues, zoning fights, and local rate backlash. That matters because the market often capitalizes announced capacity as if it were delivered capacity; if approvals slip by quarters, the revenue bridge for AMZN/ORCL becomes a timing issue rather than a demand issue. The first falsifier is a broadening of state-level incentive pauses or utility pushback that delays large loads faster than developers can re-route them.

The contrarian point is that consensus is underpricing concentration. AI buildout is likely to favor a handful of power-rich, shovel-ready corridors, which should widen the gap between asset owners and everyone else; scarcity rents may accrue to site developers, transmission, and generation owners before they show up in cloud margins. If electricity prices rise materially or public opposition expands, the market may rotate from “AI beneficiaries” broadly into a narrower basket of infrastructure enablers.

Key Decisions for Investors

• Long BX on a 3-5% pullback; 6-18 month thesis is scarcity rent capture from AI site aggregation and development fees. Falsify if major data-center projects are delayed or fee-related earnings fail to inflect.
• Hold AMZN as a conditional long only on weakness; treat this as a capacity-securement trade, not a demand trade. Best risk/reward is 12 months out, but near-term FCF compression from buildout capex is the key downside.
• Use ORCL call spreads 6-12 months out only if management confirms power-secured capacity additions; otherwise stay on watch. The main risk is bookings-to-revenue slippage from infrastructure bottlenecks, not a collapse in AI demand.
• Set a policy-risk alert on AZ/IL/VA-style incentive pauses and rate cases; if more states cap data-center benefits or impose load restrictions, rotate away from pure AI revenue names and toward infrastructure owners.