The mechanism is a rate class. On July 8, Portland General Electric began billing data centers and other new large loads under a customer class that did not previously exist, at prices roughly 30 percent higher than they had been paying. Rates for every other class of customer went down [1].
That pairing is the entire point, and it is unusual. Across most of the country, the wires, substations and generation built to serve a new data center enter the general rate base, where the cost is recovered from all customers in proportion to their use. Households capitalize the buildout and receive nothing from it. Oregon's regulators pulled the two apart.
The Oregon Public Utility Commission approved the New Large Load Tariff in May 2026, and the new prices took effect July 8 [1]. PGE describes the instrument as Schedule 96, a distinct rate class for large-load data centers [2].
The price is only half the design. The tariff carries exit fees and minimum charges, which PGE says "ensure that data centers commit to a certain level of payment, protecting families and small businesses from the costs of stranded assets" [2]. That second half may matter more than the first. A 30 percent premium does a household no good if the data center leaves in year four and the substation built for it is still sitting in the rate base. Minimum-take obligations park that risk on the customer who created it.
PGE's second-quarter results quantify the load the tariff was written against. Industrial demand rose 11.2 percent year over year, driven by high-tech and data-center expansion, while residential and commercial loads were relatively flat against the prior year [1]. Growth on that system is not coming from population or from weather. It is coming from one class of customer, which now pays for it.
The quarter itself does not yet register the change. PGE's second quarter closed June 30 and the new prices began July 8, so none of the revenue in these results reflects the tariff. The company reported $814 million in revenue against $807 million a year earlier, net income of $68 million against $62 million, and earnings of $0.59 per diluted share against $0.56 [1]. Revenue rose 0.9 percent, net income rose 9.7 percent, and earnings per share rose 5.4 percent. PGE reaffirmed full-year adjusted guidance of $3.33 to $3.53 per diluted share and serves approximately 960,000 retail customers [1].
Maria Pope, PGE's president and chief executive, framed the change as a response to cost pressure rather than to capacity or climate: "Affordability remains a national focus, and we have taken proactive steps to address customer cost pressures while supporting continued economic growth in our region" [1]. John McFarland, the company's chief customer officer, put it as an allocation question: "As energy demand grows, it is critical that the costs of new infrastructure are allocated fairly and transparently" [2].
The strongest objection to this design deserves a straight answer rather than a dismissal. A 30 percent premium is a competitive disadvantage. Data centers site on power price and power availability, and a state that charges more can lose projects to a state that does not. Losing a project is a genuine cost, not an imaginary one: construction employment, property tax base, and the load itself, which would have spread fixed system costs across more kilowatt-hours and lowered the per-unit charge for everybody.
The answer is that the premium is not a tax, it is a price. If a data center will not pay the cost of the infrastructure built to serve it, that cost does not evaporate. It is paid by roughly 960,000 retail customers who did not ask for it and cannot decline it. A project that pencils only when households finance its substation was never worth what it appeared to be worth, and the jobs it brings are partly paid for by a transfer from the people who do not get them. The exit fees test the same proposition across time rather than across state lines: a company willing to commit to a minimum payment is telling you the project is real.
The transportable part of this is a sequence, not a slogan. First, a separate rate class defined by load size, so that very large new customers stop being averaged in with households. Second, a cost-allocation proceeding that assigns the incremental infrastructure to that class instead of to the general rate base. Third, minimum charges and exit fees, so the class cannot walk away and leave the asset stranded. Oregon's version of those three steps is the New Large Load Tariff and Schedule 96 [1][2]. The names differ by jurisdiction. The sequence does not.
What came out the other end is a sentence most utility rate filings cannot produce: one class of customers received an increase of about 30 percent, every other class received a decrease, and both happened on the same day under the same order [1].
