What Oregon decided
Pacific Power has agreed to a deal requiring new data centers to bear the full cost of the energy infrastructure needed to serve them, covering new power plants, battery storage, transmission lines, substations, and connecting wires. Data centers that share infrastructure with other large customers will pay proportionally based on how much of that shared capacity their demand accounts for. The Oregon Public Utility Commission is expected to formally approve the arrangement on November 13, 2026, and Pacific Power has committed to addressing how the same cost principles apply to its existing data center customers by spring 2027, closing what had been an open question for facilities already connected.
Citizens' Utility Board spokesperson Charlotte Shuff called the agreement the strongest cost-allocation deal from a for-profit electric utility yet reached under Oregon's POWER Act, the state law aimed at protecting residential ratepayers from bearing the cost of industrial-scale demand growth. That framing matters because it signals regulators are treating this as a floor for future negotiations rather than a one-off settlement specific to Pacific Power's territory. Other state utility commissions grappling with the same politically charged question of who pays for AI-driven load growth now have a concrete, adopted structure to reference rather than a theoretical proposal.
The mechanics of proportional cost allocation
The proportional allocation model is the deal's most exportable feature. Rather than spreading the cost of new transmission and generation across the entire ratepayer base, as has traditionally happened with large industrial customers, Pacific Power will assign costs to data centers based on their share of the demand driving the investment. CUB executive director Bob Jenks illustrated the logic using rural Madras, where new infrastructure investment is, in his words, pretty clearly tied to serving data center demand rather than general population growth, making the attribution straightforward in ways it would not be in a dense urban service territory.
Transmission lines alone cost several million dollars per mile, and the Western United States needs more than 12,000 additional miles to keep pace with growing electricity demand, much of it driven by data centers and industrial electrification. Under the old cost-sharing model, that expense would have been socialized across residential and commercial customers who see none of the economic benefit from a given data center campus. Under the new structure, the entity creating the demand pays for the infrastructure that demand requires, a principle simple enough to explain to a state legislature and specific enough to survive a rate case.
Winners and losers across utility territories
The deal creates an immediate competitive divide within Oregon itself. Facilities in Portland General Electric's service territory, where smaller data centers benefit from costs spread across a broader customer base, retain a cost advantage that Pacific Power's territory no longer offers to new entrants. That gap is not hypothetical. The article reporting on the deal notes dozens of large facilities are currently waiting to connect to Pacific Power's system, and each one now faces a materially different cost profile than it would have under the prior framework, a change significant enough to affect site-selection decisions already in motion.
This kind of intra-state divergence is likely to become a recurring feature of data center site selection rather than an Oregon-specific quirk. As more utilities adopt proportional cost allocation in response to regulatory and political pressure, the patchwork of rules across neighboring service territories, and even across neighboring states, will widen. A site that looked economically comparable to a competing location eighteen months ago can now carry a meaningfully different total cost of ownership once the local utility's cost-allocation policy is factored in, which raises the value of running that analysis early rather than after a site is already under contract.
Industry pushback and the road ahead
The Data Center Coalition, the industry's primary advocacy group, is expected to appeal the decision, arguing that the cost burden could discourage investment in a state competing for data center projects against jurisdictions with less restrictive rules. That tension is real: Oregon risks pushing marginal projects toward states with looser cost-allocation regimes, even as it protects existing residential ratepayers from subsidizing that growth. How regulators balance those competing pressures over the coming year will shape whether other states move quickly to adopt similar rules or hold back to avoid losing projects to more permissive neighbors.
The spring 2027 deadline for addressing existing data center customers is the next milestone worth tracking closely. Facilities that signed connection agreements under the old cost-sharing framework will want clarity on whether new rules apply retroactively or only to future expansions, and that answer will set precedent for how utilities elsewhere handle the transition for their own existing large customers. Any operator with active facilities in Pacific Power's territory should be tracking that proceeding directly rather than waiting for the outcome to be reported after the fact, given how much it could reshape ongoing operating costs.
Why this becomes the national template
Other states have also grappled with data center cost allocation, and this deal stands out for being negotiated and adopted rather than merely proposed or litigated. That distinction gives other state utility commissions, and the utilities themselves, a working example to cite when residential and commercial customer groups push back on rate increases tied to industrial load growth. Expect utility filings in other states over the next two to three quarters to reference the Pacific Power framework explicitly, particularly the proportional allocation mechanic, since it offers a politically defensible middle ground between blocking data center growth outright and asking existing ratepayers to fund it.
The broader trend this reflects is the end of an era where data center operators could treat grid interconnection costs as a shared utility expense outside their own budget. As AI-driven demand pushes utilities toward historic levels of new generation and transmission spending, state regulators face direct political pressure from residential customers seeing rate increases tied to industrial growth they do not benefit from. Cost-causation principles like Oregon's are the release valve for that pressure, and operators should expect the direction of travel across most states to move toward, not away from, this kind of allocation over the next several years.
What this means for infrastructure roadmaps
For any enterprise or colocation operator planning new data center capacity, the immediate takeaway is that utility-specific cost allocation policy needs to move up the site-selection checklist to sit alongside power availability, latency, and tax incentives. A location with abundant power but an unfavorable cost-allocation regime can end up more expensive over a facility's operating life than a location with less available capacity but a friendlier cost structure, and that comparison only becomes visible if the analysis is run explicitly rather than assumed to be similar across states.
Procurement and infrastructure teams negotiating colocation or build-to-suit agreements should also push for contract language that addresses how future changes in utility cost-allocation policy get passed through, since Oregon's shift shows these rules can change materially within a single planning cycle. A facility signed today under one cost regime could face a different one within eighteen months if a state regulator adopts a Pacific Power-style framework mid-contract. Building that contingency into contracts now is cheaper than renegotiating after a rate case has already gone against you.
