Sure, Build a Data Center in Oregon, But Pay the Cost.
Editors note: This is an updated version of this essay. Good friends involved in energy and environmental policy spent time giving me critique, evidence and nuance against a relatively generic essay. I am grateful to them, and the citizens making it clear how important a priority this is to their political choices. I’ve updated the essay to include more and will continue to refine my approach and be clearer and more concrete on how these issues manifest for Oregonians. This is what I mean by making campaign thinking public and supporting a more complex public discourse. Thank you for reading and engaging.
Oregonians are being told a familiar story.
Energy costs are rising because of climate policy. Because of population growth. Because of the transition to renewables. Because of “market forces” no one can quite name.
What we are not being asked to talk about clearly is this: who is driving new demand on the grid, who benefits from it, and who is paying for the long-term costs.
Oregon is rapidly becoming a hub for energy-intensive infrastructure. Communities around Oregon are negotiating how to balance the needs of their industrial sector, attract investment for public infrastructure, and serve the competing needs of citizens in their jurisdictions. Oregon is an attractive site for developers because we have historically had relatively low electricity costs, publicly regulated utilities, abundant hydroelectric power, a political culture that is eager to attract “clean tech” investment, and rural districts hungry for revenue.
The regulations around electricity and industry made sense when the grid was sized for households, small businesses, and traditional industry. We knew how to plan for and measure against those needs. They make far less sense when we are grappling with some of the most energy-hungry operations on the planet and trying to understand where regular people will absorb the consequences.
This isn’t an argument against data centers or AI. It is an argument for a nuanced analysis of the regulatory environment that surrounds them.
Oregon's Utility Structure Is More Complex Than the Headlines
Before I make any policy argument, I want to be honest about something else I oversimplified in my first version.
Oregon's energy landscape is not monolithic. The majority of Oregon's utilities are consumer-owned cooperatives and public utility districts already accountable to their local communities by design. Many incentive structures requiring local community investment are already in place and working. The investor-owned utilities most Oregonians are familiar with as residents of urban areas serve under a different accountability structure, regulated by the Public Utility Commission with obligations to shareholders as well as ratepayers. There are real issues in incentives and accountability mechanisms in that approach and I think its worth talking about that too.
The real pressure points in Salem right now are transmission and resource adequacy- the capacity of the broader grid to move power where it needs to go, and Oregon's ability to reliably meet peak demand as large new loads come online simultaneously. Those are statewide infrastructure questions that require statewide coordination, and the public needs to understand the tradeoffs more clearly to move beyond another reductive culture war.
Two Oregon Communities, Two Different Stories
The contrast between Morrow County and Hermiston illustrates why this conversation is hard and complicated and why the for/against framing fails us in this moment.
Morrow County welcomed Amazon's first hyperscale data center in 2011. Amazon has since constructed seven facilities in the area, with agreements for five more underway, generating commercial taxes worth more than $100 million over a decade, providing real economic value for a rural county that had lost timber jobs and a coal plant.
But Morrow County is also dealing with a serious water crisis and it's confounding that the port is ready to supply water for Amazon's exascale data center while some residents don't have clean drinking water. It’s also clear that infrastructure costs weren't fully accounted for when making that first deal. Upgrading the water system and reinforcing against failure points should clearly have been part of the initial negotiations and wasn’t. That gap is where we need attention and focus.
Hermiston tells a different story. In February, Hermiston and Amazon reached a 25-year agreement that would expand summertime water capacity by 100 million gallons annually for all users, thanks to innovations and investment in storing excess winter flows. Hermiston's utility rates went down because of Amazon's presence. Workers and families in that community have a relationship with this industry that I am not going to dismiss or characterize as a problem. It’s an example of a company investing in shared community infrastructure that has real benefits for real Oregonians we can’t just ignore. That's closer to a model worth building on.
In Prineville, Meta has donated more than $2 million to Crook County School District over the last decade as it built one of its largest data center complexes. That’s a real community benefit, but it’s also offset by the tax breaks offered to them to build. Also true: Meta's Prineville facility used fewer gallons per minute than other existing industrial users in the same area.
The full accounting matters. Oregon school districts lost $275 million to property tax abatements in 2024, more than double what they lost in 2019. Crook County alone lost $29 million in a single year. Meta's $2 million in school donations over a decade, while genuinely appreciated, does not close that gap. Philanthropy is not a substitute for a tax base. If we want data centers to be genuine community partners, the accounting has to be honest about both sides of the ledger.
The question isn't which story is true. Both are true. The question is what model serves Oregon best as we think about development projects of this kind.
Information and Leverage Are the Policy
Here's what I think Hermiston did better than Morrow County: they negotiated from a position of better information and with clearer expectations about what the community required in return.
That points directly to what a Governor can actually do.
Oregon can organize that effort from Salem so that rural communities aren't negotiating alone. A statewide framework for data center accountability, standardized disclosure requirements, model agreements, clear expectations about community benefit, means Boardman and Hermiston and the next community in line aren't each reinventing the wheel against some of the most sophisticated corporate negotiators in the world.
On water specifically: the prior appropriation doctrine cuts both ways. If Oregonians own the water, there are mechanisms to require compensation when corporations draw from public aquifers. Beneficial use determinations applied to data centers drawing from public water sources deserve serious examination. Water use disclosure requirements (which some companies have actively resisted) should not be optional. The POWER Act is a start. There is work to do in Salem on water transparency amendments that deal with downstream impacts more directly, and I would be glad to work with Democrats there to develop them.
How we manage both water and power as resources will likely be the defining fights of the next decade in Oregon. I’m proud to fall on the environmental conservation end of the scale, but I recognize that developments of this kind offer real opportunities to communities considering them.
When I was advocating for cannabis, the insanity of starting water-intensive agricultural crops at scale in states like Nevada and Arizona who already hurt for water, instead of common sense interstate commerce regulation for trade was an important point I made at both state and federal level and is an example of the ways these issues cut across industry and will require systems-level analysis and creative thinking to find paths forward.
What Honest Accounting Looks Like
The tools already exist. They're just not being applied consistently, and local communities are negotiating on an extremely uneven playing field.
Higher marginal energy rates for large continuous loads. Grid reinforcement fees and other infrastructure safety and redundancy paid by the projects that require them. Long-term infrastructure depreciation charges that assign responsibility for accelerating wear. Water infrastructure investment as a condition of approval, not a philanthropic gesture after the fact. Community innovation investment formalized as a standard expectation rather than a negotiated exception.
The Hermiston water storage agreement is a model for what depreciation thinking looks like in practice. Amazon messed up in Boardman and came back to the table with a more honest accounting of its impact on a shared resource in Hermiston, and invested in expanding that resource for everyone. That's not charity. That's honest accounting.
What I would ask companies who want to build here: pay the true marginal cost of energy, contribute to the grid upgrades you require, invest in local water infrastructure, and be transparent about your resource consumption. That doesn't drive investment away. It attracts the kind that plans to stay.
The Choice We're Actually Making
Every energy policy is a moral choice, whether we name it or not.
We can continue to socialize industrial costs, privatize profits, and tell households to absorb the difference. Or we can say, clearly and calmly: you can build here. You just have to pay for what you use, invest in the communities that host you, and leave the infrastructure in better shape than you found it.
That is not hostility to data centers or to AI. It is a requirement that development serve Oregon's whole community - not just the portions of it with the leverage to negotiate good deals on their own. It is also a willingness to accept the limitations we have to constrain the ongoing development underpinning the AI push from the tech sector. Oregon is ready to do the math, we just need to make it more transparent.