Oregon regulators move to make data centers pay more for grid expansion

Oregon regulators issued the first major order under the 2025 POWER Act, creating Schedule 96 to require large data centers to pay more directly for grid expansion costs. The framework adds long-term contracts, minimum payment obligations, exit fees, a one-cent-per-kWh surcharge for some 100 MW+ users, and a peak growth modifier to shift more transmission and generation costs to heavy-load customers. PGE still must file specific rate changes, with implementation expected to begin June 10, and similar proceedings are likely for Pacific Power.

Analysis

This is a structural margin reset for regulated utilities in fast-growth territories: the key change is not higher near-term revenue, but lower probability that load growth turns into unrecoverable capex dilution. That improves the quality of incremental earnings for POR because the utility can pursue interconnection growth without fully socializing grid expansion, which should reduce political/regulatory overhang on future rate cases. The second-order winner is the entire “picks and shovels” ecosystem tied to dedicated load buildouts—substation equipment, switchgear, transformers, and EPC capacity—because customers now have stronger incentives to fund bespoke infrastructure instead of waiting on general system upgrades.

The loser set is more interesting than just data centers: hyperscalers lose optionality, and the economics of speculative site hoarding get worse. Developers that rely on cheap, utility-funded transmission runway will face longer contracting cycles, higher deposits, and stricter take-or-pay terms, which likely shifts bargaining power back toward utilities in other Western states over the next 6-18 months. This also creates a subtle anti-growth risk for local commercial real estate and municipal tax base assumptions if some projects are delayed, resized, or redirected to jurisdictions with more permissive tariff treatment.

The market may be underestimating how limited the direct benefit to residential bills will be in the first 1-2 years. The headline is “data centers pay,” but the real transmission path is slower: it reduces future cost-shifting and lowers the probability of a stranded-capex cliff, rather than mechanically cutting existing bills. That means the strongest near-term trade is not a consumer relief basket, but a relative de-risking of regulated utility valuation versus unpriced AI-load uncertainty.

Catalyst-wise, the next leg comes from Pacific Power’s proceeding and from PGE’s actual tariff filing, where the real economic transfer will show up. If those filings land materially above current expectations, the sector could see a broader repricing of large-load interconnection assumptions across the Northwest within one quarter. Conversely, if the surcharge is largely symbolic or softened in implementation, the move fades and POR gives back most of the regulatory premium quickly.

Key Decisions for Investors

• Long POR on a 3-6 month horizon: this reduces tail risk around future capex socialization; best entry on any post-news pullback, with upside from multiple stability rather than earnings beats.
• Pair trade: long POR / short a hyperscaler proxy or data-center-enabling REIT basket over 6-12 months to express the margin transfer from utility customers to dedicated-load customers.
• Buy regulated utility peers with pending large-load cases on dips, but underweight names with heavy AI-load exposure until their tariff frameworks are clarified; risk/reward improves only after explicit cost recovery language.
• Watch for a follow-on long in industrial infrastructure names tied to grid expansion if Pacific Power mirrors this structure; the cleanest trade is via contractors and electrical equipment suppliers with 12-18 month backlog visibility.
• Avoid chasing data-center landlord exposures until the market prices in higher deposits, exit fees, and minimum-take requirements; those terms can compress IRRs by 100-200bps if replicated broadly.