bytevyte
bytevyte
Language
ai-beats —

Oregon Data Center Power Demand Reaches 23% of Retail Electricity, and Regulators Have Started Billing for It

Oregon data center power demand

Oregon's data centers drew roughly 23% of the state's retail electricity in 2025, and state regulators have now put a price on that concentration. The Oregon Public Utility Commission approved Schedule 96 for Portland General Electric under the state's POWER Act, adding a 1 cent per kilowatt-hour premium to large-load customers at or above 100 megawatts. Oregon data center power demand is the figure that made the tariff hard to avoid: 111 operating facilities, 22.1 million square feet of floor space, and 32 more sites already in development, according to a statewide industry inventory.

The inventory, published on September 17, 2026, is the first attempt to count the sector's footprint in Oregon. It puts the fleet at 111 sites and about 2,630 direct jobs, and lists 32 projects under development.

Hyperscale capacity clusters in Eastern Oregon, where data centers supply more than 10% of the workforce in Morrow County. The tax picture is where the arithmetic turns uncomfortable: the inventory reports $60.2 million in property taxes paid by the facilities in 2025, against estimated abatements of $230 million to $450 million.

How Oregon Data Center Power Demand Reached 23%

Run the per-unit numbers and the development bargain looks thin. Spread across 22.1 million square feet, $60.2 million in property taxes works out to about $2.72 per square foot. Per direct job, it is roughly $22,900. Set against the abatement range, the state forgoes between $3.80 and $7.50 in potential revenue for every dollar it collects.

Job density is the number that stands out. Across 111 facilities averaging about 199,000 square feet each, 2,630 direct jobs works out to roughly 24 positions per site, or one direct job per 8,400 square feet of floor space. Cooling, power management, and monitoring automation keeps that ratio low. The 32-site pipeline is a 29% increase in facility count, and the employment gain scales with it rather than with the electricity drawn.

Concentration is the risk attached to the Eastern Oregon cluster. When one industry supplies more than a tenth of a county's employment, the county budget and its services track that industry's construction and pause cycles. Morrow County has little diversification against a hyperscale slowdown, and the abatements defer revenue while the load and its grid costs arrive first.

Electricity is the constraint that arrives first, and the load curve is the hard part. Data centers run flat rather than following residential demand patterns. Oregon data center power demand is a question of shape as much as volume, because the sector's draw does not fall off at night the way household demand does. Interconnection queues and transmission upgrades are the bottlenecks, and those timelines are measured in years.

Inside the One-Cent Large-Load Premium

Schedule 96 is narrower than its reputation suggests. It applies only to customers at or above 100 megawatts, which means the tariff covers hyperscale campuses and leaves smaller colocation and edge sites untouched. The premium is 1 cent per kilowatt-hour on top of standard rates.

The arithmetic is simple. A 100 megawatt facility running at an 80% capacity factor consumes roughly 700 gigawatt-hours a year, so the premium adds about $7 million to its annual bill at the threshold. For a campus with capital costs in the billions, that is a rounding error in one line item.

The threshold is where the design gets shaky. A premium that starts at 100 megawatts covers today's hyperscale campuses and exempts everything below. Any load under that line, including regional colocation and smaller inference clusters, keeps paying standard rates. How much of the sector pays the premium depends on how the fleet is built, not on how much power it consumes in total.

The tariff's purpose is to shift grid costs onto the load that creates them, instead of spreading those costs across residential and commercial ratepayers through general rates. Whether it accomplishes that depends on how many loads cross the 100-megawatt line, and that is a function of how the next round of campuses is built.

Oregon is one of 24 states that have adopted independent large-load standards or tariffs. Oregon's version applies only above 100 megawatts, so the facilities below that line keep paying standard rates.

The Counter-Argument Worth Taking Seriously

The case against Oregon's premium is that it changes nothing. Thirty-two facilities remain in development despite the tariff, and hyperscale siting turns on land, fiber routes, water, and interconnection speed far more than on a cent per kilowatt-hour. If operators are price-insensitive at this margin, the premium raises revenue without changing behavior.

That argument is right about the cent and wrong about the direction. The revenue is real, and it moves a slice of grid cost recovery onto the load that creates it. The weak point is the assumption that price is the only lever. A one-cent charge on consumption does not make a speculative capacity reservation expensive, and it does not give the grid authority to cut a load during stress. Treating it as the end of the policy response would be a mistake.

The federal layer is stalled. The Ratepayer Protection Act and the GRID Savings Act remain stuck in the Senate, and no further federal action is expected before the November 2026 midterms. The next round of state filings will settle what the template for the sector looks like.

Why this matters

The Oregon inventory is the first hard measurement of what a data center boom does to a state power system: 23% of retail electricity from 111 facilities, about 2,630 jobs, and $60.2 million in property taxes against abatements worth several times that. The one-cent premium applies to a narrow slice of the load and leaves most of the sector on standard rates. Grid capacity and transmission timelines, rather than the size of the premium, will decide where the next campus lands.

✔Human Verified


Researched and cross-referenced against primary sources by the Bytevyte editorial team. This article was generated with the assistance of artificial intelligence and reviewed by the Bytevyte editorial team.