Data centers have already hiked electricity prices on the public by $23 billion. Good luck clawing that back

PJM data indicates expected power demand from data centers is driving $23B in customer electricity price increases lasting through at least 2028. The article highlights regulatory complexity in allocating grid-investment and transmission/substation costs—creating uncertainty for ratepayers and potential “loopholes” where data centers could reduce charges tied to coincident peak demand. With rate impacts likely unfolding over years and representation gaps for residential consumers, near-term consumer pressure and regulatory scrutiny could rise.

Analysis

This is a slow-burn tariff and rate-base story, not a near-term earnings shock. The market error is to assume AI load growth automatically accrues to utilities; in practice, the winners are the balance sheets that can get substations, transmission, and interconnect costs into rate base quickly without triggering a political backlash. If commissions socialize too much of the spend, the margin hit lands on households and small commercial users, which raises the odds of headline-driven rate caps and multiple compression in the highest-growth utility territories.

The cleaner second-order beneficiary is the grid buildout complex: ETN, PWR, GEV, and similar EPC/equipment names should see multi-quarter order support even if some data-center projects are delayed, because utilities still need the hardware once planning is approved. The more exposed names are the marginal data-center operators and landlords whose leasing economics depend on all-in power costs staying stable; if large-load tariffs reset higher, weaker projects in PJM-heavy markets become uneconomic first.

Contrarian view: consensus focuses on "who pays," but the real edge is who can game the methodology. Sophisticated large loads can dodge coincident-peak charges, so the system may still undercharge data centers while overcharging everyone else. That makes the base case less about one clean transfer and more about a patchwork of state rulings, with the most important catalysts arriving over 1-3 months in docket decisions and over 6-18 months in actual rate cases and project cancellations.

Key Decisions for Investors

• Favor a 6-12 month long basket of ETN / PWR / GEV on any pullback: these names monetize the infrastructure spend regardless of whether utilities win the political argument. Risk/reward is attractive if PJM-style load growth keeps capex elevated; thesis breaks if large-load projects get canceled faster than utilities can order equipment.
• Do not short the hyperscaler basket (MSFT / AMZN / GOOGL / META) on this headline alone. The power-cost impact is more likely to show up as project timing noise than as a durable change in free-cash-flow generation; use only as a hedge if multiple states adopt explicit large-load surcharges.
• If state commissions in PJM-heavy markets approve explicit large-load tariffs over the next 1-3 months, rotate into regulated utilities with rate-base growth visibility and away from politically exposed names. Best expression is long XLU only after a selloff caused by consumer-rate backlash; stop if rate-case language starts capping recovery.
• Watch EQIX / DLR as a conditional short only if tenants cannot fully pass through power costs or if lease spreads roll over. This is a lower-conviction trade until we see actual occupancy or renewal-rate pressure; otherwise the stocks can absorb higher utility bills.