Back to News
Market Impact: 0.45

How families could get stuck with higher electric bills if the AI data center boom goes bust

CEGVSTAEP
Artificial IntelligenceEnergy Markets & PricesRegulation & LegislationInfrastructure & DefenseTechnology & InnovationConsumer Demand & RetailLegal & LitigationInvestor Sentiment & Positioning
How families could get stuck with higher electric bills if the AI data center boom goes bust

PJM customers face $16.6 billion in charges to secure future power for 2025–2027, with Monitoring Analytics attributing about $15 billion (≈90%) of that to projected data center demand; PJM forecasts ~30 GW of added load from data centers through 2030. Retail electricity prices have risen sharply year-over-year in key data-center states (September: Illinois +20%, Ohio +12%, Virginia +9%), and utilities are tightening interconnection rules—AEP cut Ohio data-center connection requests from 30 GW to 13 GW after imposing 85% payment and exit-fee requirements. Monitoring Analytics has filed a FERC complaint urging PJM to deny connections unless supply is secured, highlighting regulatory and stranded-cost risk for utilities, consumers and data-center developers.

Analysis

Market structure: Data-center capex is transferring near-term utility/capacity risk to ratepayers — winners are regulated utilities able to recover transmission capex via rate base (e.g., AEP) and transmission contractors; losers are merchant generators (Vistra/VST, uncontracted CEG capacity) and residential ratepayers. The 30GW PJM forecast (vs. AEP's 13GW after rule changes) implies large upside for transmission build but high uncertainty — expect project-level churn and site-shopping to compress merchant pricing power and increase short-term capacity-market volatility. Risk assessment: Tail risks include FERC/state rulings that either (1) force stricter interconnection (reducing forecasted load by >30%) or (2) mandate socialization of stranded costs — both could re-rate merchant generators and utilities differently. Immediate catalysts (days–weeks): Monitoring Analytics FERC complaint and PJM queue revisions; short-term (3–12 months): state commission adoption of AEP-style binding payment/exit fees; long-term (2–5 years): actual buildout vs. forecast determines stranded assets vs. regulated rate base recovery. Trade implications: Favor regulated utility exposure with constructive regulatory clarity (AEP/AEP) and underweight merchant, commodity-exposed generators (VST), specifically via 6–12 month directional positions and options to express convexity. Use pair trades (long AEP, short VST) to isolate regulatory/revenue recovery vs. merchant demand risk; lean into volatility in capacity/wholesale markets with put buys on VST (6–12m) and call spreads on AEP (9–12m) for asymmetric payoffs. Contrarian angles: Consensus assumes large net new load; reality may be a geography-shift + self-generation (onsite gas+storage) outcome that leaves fewer grid customers paying stranded costs — this would sharply hurt merchant names and benefit battery/onsite generation vendors. The market may be underpricing regulatory intervention risk: a FERC order forcing data centers to prove committed load would be a >30% downside catalyst for speculative-build beneficiaries and a lift for disciplined regulated utilities.