Fact Check Team: Should data centers foot the bill of new power generation?

U.S. residential electricity prices are at decade highs and NEADA estimates average household winter heating costs of $995, up $84 year-over-year, with average monthly bills up ~10% since January. The Biden-era statistic is cited as the administration (via DOE and the National Energy Dominance Council) is pushing PJM to accelerate more than $15 billion of new “reliable baseload” generation (coal, natural gas, nuclear) in the Mid-Atlantic and to require large new electricity users — notably data centers — to shoulder buildout costs rather than ratepayers. The proposal could reallocate financing risk to data-center operators, affect regional utilities and project financing amid high interest rates and aging infrastructure, and has implications for consumer bills and energy-sector regulation.

Analysis

Market structure: Policy pushing PJM toward $15B+ “reliable baseload” favors incumbent regulated utilities and large generation owners (Exelon, Duke, NRG) who can recover regulated or contract revenue; losers are data‑center landlords/operators (DLR, EQIX) and hyperscalers if costs are socialized to them, compressing FCF and growth. Expect near‑term supply tightness (0–24 months) as permitting/finance delays keep marginal capacity low, pushing spot power and natural gas prices up 10–25% in stress episodes; credit spreads for merchant generators and muni/utility debt should widen if rates remain elevated.

Risk assessment: Tail risks include federal litigation or state PSC pushback that nullifies DOE cost allocation (low probability, high impact), or a tech-led relocation of capacity outside PJM reducing demand (medium probability). Immediate (days–weeks): market repricing on announcements; short (3–12 months): regulatory rulemaking and contract negotiations; long (2–5 years): actual buildout and transmission upgrades — the latter is the biggest hidden dependency and cost driver. Catalysts to watch: DOE/PJM formal rule, state PSC orders, and quarterly commentary from EQIX/DLR within next 90 days.

Trade implications: Direct plays — overweight regulated utilities (EXC, DUK) and gas producers (XOM) for 6–12 months; short/hedge data‑center REITs (DLR, EQIX) tactically for 3–9 months. Options: buy 6–9 month calls on EXC (10–15% OTM) and buy 3‑month put spreads on DLR (10%/20% OTM) to cap cost. Rotate 1–3% portfolio from hyperscalers (AMZN/MSFT/GOOGL) into utilities/gas to hedge rising grid costs.

Contrarian angles: Consensus that data centers will fully shoulder costs misses that many hyperscalers have long‑dated contracts and bilateral PPAs — this limits near‑term pass‑through and could insulate REIT cashflows, making current selloffs overdone. Historical parallel: telecom cost allocation battles delayed investment but ultimately created contracted capacity markets; unintended consequences include faster adoption of on‑site generation or edge computing, benefiting small modular generator builders and industrial power players.

Key Decisions for Investors

• Establish a 2–3% long position split between Exelon (EXC) and Duke Energy (DUK) (1–1.5% each) targeting a 6–12 month hold; add another 2% if DOE/PJM issues formal binding rules within 90 days; use a 20% trailing stop.
• Initiate a 1–2% short position in Digital Realty (DLR) or Equinix (EQIX) (or a pair trade: long EXC 2% / short DLR 2%) for a 3–9 month horizon to capture margin compression risk; hedge with bought 3‑month put spreads on DLR (buy 1× put 10% OTM, sell 1× put 20% OTM) to cap downside risk.
• Allocate 1% to buy 6–9 month call options on NRG Energy (NRG) or EXC ~10–15% OTM as a leveraged play on capacity buildout and merchant-to-contracted revenue re‑rating; max allocation per name 1% to limit theta decay risk.
• Reduce exposure to hyperscalers (AMZN, MSFT, GOOGL) by 1–2% and redeploy into integrated energy producers (XOM or CVX) 1–2% for 6–12 months to hedge upside in natural gas-driven power costs; reassess after quarterly earnings and any state PSC rulings within 90 days.