Boardman / Umatilla Data Centers and Infrastructure

What the public record shows about data center growth, electricity infrastructure, agriculture, and the sequence of development in northeastern Oregon.

What we know so far

The power piece is actually one of the easier parts to follow.PDX-4 was operating by 2013, but BPA records show power requests tied to PDX-1 and PDX-4 going back to 2011. Then in 2012 there was a much bigger request, around 225 MW, and Morrow Flat was already listed as the point of interconnection.Then BPA publicly proposed Morrow Flat in 2013.What I think is interesting is that BPA did not describe that project as being for one customer. They said it was meant to help serve growing demand from new data centers, food processing plants and existing development.So data centers are clearly part of why the infrastructure was expanding, but that still does not tell me how much of it was specifically for Amazon or who paid for what.The Morrow Flat substation also used about 22 acres of land that had previously been used for agriculture, which is not huge in the grand scheme of Boardman, but it is still worth noting because this is all happening inside an agricultural region.The water piece gets a lot fuzzier.I have confirmed two AWS-associated drinking water systems that use groundwater. What I still cannot get my hands on are the actual facility-level numbers. How much are they withdrawing? How much are they consuming? How much is going back out as wastewater?And that leads into the Port of Morrow system, because the Port takes industrial wastewater from data centers and land applies industrial wastewater to agricultural land.That sounds simple until you try to trace it.There are different facility names, different wastewater source names, different permits, and I am still trying to figure out exactly which data center connects to which source and where that water actually ends up.The nitrate issue is another place where I had to change how I was looking at Boardman.The Lower Umatilla Basin was designated as a groundwater management area in 1990, so the nitrate problem was already well established long before the major data center buildout.That does not make data centers irrelevant to the water question. It just changes the question.Instead of asking whether data centers caused Boardman’s groundwater problems, I think the more useful question is what happens when you add this much new industrial demand into a place that already had pressure on groundwater, wastewater systems, farmland and infrastructure.That is where Boardman starts getting really interesting.

Infrastructure

Boardman already had a lot going on before the major data center buildout. Agriculture, food processing, power infrastructure, groundwater use, wastewater systems and industrial development were all already part of the picture.Then data centers started growing into that existing system.So the question is not just what got built for data centers. It is what was already there, what had to expand, what changed as demand grew and who paid for those changes.ElectricityThe power story starts before PDX-4 was operating.BPA records show a request tied to PDX-1 and PDX-4 in 2011 for about 60 MW. Then in 2012 another request tied to PDX-1 and PDX-4 came in at about 225 MW, with Morrow Flat listed as the point of interconnection.PDX-4 was operating by 2013.Later that year BPA publicly proposed the Morrow Flat project. BPA said the project would help serve up to 250 MW of growing demand from new data centers, food processing plants and existing development.So data centers were clearly part of the reason the electrical system was expanding.What I still cannot tell you is how much of Morrow Flat was specifically tied to PDX-1 and PDX-4, whether those loads directly triggered the project, or exactly how the costs were divided.We also know PDX-4 had to be getting power somehow before Morrow Flat was completed. BPA records show existing 115-kV delivery infrastructure serving the Boardman area before PDX-4 began operating.That leaves one of the questions I am still chasing: what exactly supplied PDX-4 at the beginning, and how did that power path change as the data centers grew?Water and wastewaterWater is harder to follow.I have confirmed two AWS-associated drinking water systems in the region that use groundwater. What I still do not have are good facility-level numbers showing how much water those sites are actually withdrawing or consuming.The wastewater side adds another layer.The Port of Morrow operates a large industrial wastewater system that takes wastewater from several types of industrial users, including data centers. Port and DEQ records identify data center cooling water wastewater as part of that system.That wastewater is also land applied to agricultural land.One of the sources identified in the records is called VA Data #4. I know it is tied to a data center, but I still have not been able to confidently connect that name to a specific AWS facility.I am also still working on how much wastewater comes from that source, what is in it, where it moves through the Port system and which agricultural fields ultimately receive it.And that is important because simply knowing that data center wastewater enters the Port system does not tell us what effect, if any, it has on groundwater or agriculture.LandThere is a physical land piece too.BPA records show the Morrow Flat substation used about 22 acres of land that had previously been used for agricultural production.BPA also said it tried to place transmission infrastructure in ways that reduced impacts to agriculture and that landowners would be compensated when impacts could not be avoided.Twenty-two acres by itself does not tell us that data center growth has meaningfully reduced agricultural production in the region.But it is a documented example of agricultural land being converted for infrastructure connected to regional industrial growth, including data centers.And as more campuses, substations, transmission lines and related infrastructure are built, I think it is worth keeping track of how much agricultural land is actually being converted and where.Why the picture is hard to pin downThe frustrating part is that none of this lives in one place.BPA records tell me part of the electricity story. DEQ records tell me part of the water and wastewater story. The Port of Morrow has another piece. County planning records show land use and development. Utility agreements can answer parts of the cost question.But the thing I usually want to know is the connection between all of them.Which facility needed what?What was already there?What had to be expanded?Who paid for it?And what did those changes mean for the farms and communities that were already using the same systems?That is the part that is much harder to find in a single public record.It does not automatically mean somebody is hiding something. It does mean that understanding what actually happened takes pulling records from a bunch of different places and putting the timeline back together.
And that is basically what I am doing here.

The Boardman case is less about a single facility than about the network built around regional growth.Existing BPA records show that Boardman already had 115-kV delivery infrastructure before PDX-4 began operating. Later records document the Morrow Flat expansion as demand increased from data centers, food-processing plants and other development.What is still missing is the detailed map between those facts: which substations and feeders served the data center at each stage, what upgrades were triggered as load increased, and which costs were assigned to individual customers versus the broader utility system.Those details matter because infrastructure decisions can outlast the projects that initially drove them. This case study is therefore tracking not just whether new electrical facilities were built, but how responsibility for those investments was allocated and who ultimately carried the risk.

What happened when

1990
The Lower Umatilla Basin was designated as a Groundwater Management Area because of elevated nitrate levels.That is important context because the regional groundwater problem was already well established long before the major data center buildout.2011
BPA records show a roughly 60 MW load request called “PDX1&4 Data Centers.”Boardman also already had Umatilla Electric Cooperative 115-kV delivery infrastructure in place at this point.2012
A second BPA request, “PDX-1 and PDX-4 Data Centers, Phase 2,” entered the queue at about 225 MW.This one listed Morrow Flat as the point of interconnection.That caught my attention because Morrow Flat had not even been publicly proposed yet.2013
Amazon’s PDX-4 data center was operating in Boardman.So by this point, a large data center was already running while the regional power system was still changing around it.November 2013
BPA publicly proposed the Morrow Flat Substation and related transmission upgrades.BPA described the project as helping serve increased demand from existing development, food processing plants and new data centers in the Boardman area.2014
BPA project records described Morrow Flat and the related upgrades as helping serve up to 250 MW of additional Boardman-area demand.The project included a new 230/115-kV substation and other transmission work.The substation site also used about 22 acres that had previously been used for agriculture.By early 2016
Morrow Flat was in service.So the basic sequence is pretty clear: data center load requests show up first, PDX-4 starts operating, and the larger regional infrastructure expansion follows.What is still not clear is exactly how much of that later expansion was tied to specific data centers, what supplied PDX-4 before Morrow Flat was finished, and who paid for which pieces.

What the records establish

The Boardman power story starts earlier than I expected.BPA records show a request called “PDX1&4 Data Centers” entering the queue in December 2011 for about 60 MW. Then, in August 2012, another request came in for about 225 MW under “PDX-1 and PDX-4 Data Centers, Phase 2.”That second request listed Morrow Flat as the point of interconnection.The weird part is that Morrow Flat had not even been publicly proposed yet.By 2013, PDX-4 was operating in Boardman. Later that year, BPA publicly proposed the Morrow Flat Substation and related transmission upgrades. BPA said the project would help serve increased demand from new data centers, food processing plants and existing development.So by the time Morrow Flat was announced publicly, there had already been years of data center load planning happening behind the scenes.The Morrow Flat project itself was not small. BPA described a new 230/115-kV substation, transmission upgrades, new conductor, tower work, communications upgrades and road improvements. The substation site also used about 22 acres of land that had previously been used for agriculture.That is one of the things I keep coming back to with Boardman. The data center story is not just about the buildings. It is also the infrastructure that has to show up around them.The water side is less complete, but there are still a few things we can say.State records show two AWS-associated drinking water systems in the area that use groundwater. One is tied to Amazon Data Hermiston PDX080 and another to Amazon Data Umatilla PDX063.That does not tell us how much water those facilities use, which is the number I actually want. But it does tell us that at least some AWS facilities in the region are relying on groundwater.Then there is the Port of Morrow wastewater system.Port and DEQ records identify data centers as part of the industrial wastewater stream, including cooling water blowdown. That wastewater moves through the Port system and is land applied to agricultural land.One source in the records is called “VA Data #4.”That name has been driving me a little crazy because I know it is a data center source, but I still do not have a clean facility crosswalk telling me exactly which AWS site it belongs to.That is the kind of thing that keeps happening in this case. The pieces are public. The names are public. The permits are public. But figuring out how the pieces connect is a whole different job.And then there is nitrate.The Lower Umatilla Basin was designated as a Groundwater Management Area in 1990 because of elevated nitrate levels.That was more than twenty years before PDX-4 started operating.Which means Boardman did not suddenly become a groundwater problem because data centers arrived. The region already had a serious nitrate issue tied to agriculture, livestock, food processing, wastewater and other sources.What I want to know now is what happens when you add a completely new industrial load on top of all of that.More power demand. More infrastructure. More land conversion. More water demand. More wastewater.That is where the Boardman story gets a lot more interesting than just “Amazon built data centers here.”

Sources & primary records

Most of what I am using for this case study comes directly from public records.Some of the most useful records so far have come from:Bonneville Power Administration (BPA)Transmission planning, load requests, outage records and project documents are helping me piece together when electrical demand showed up and what infrastructure followed.Morrow Flat Substation and Transmission Line Improvement ProjectOregon Department of Environmental Quality (DEQ)DEQ records have helped establish when individual facilities were operating and how they were regulated.Amazon PDX-4 Data CenterDEQ records are also important for understanding the groundwater issue and the nitrate problem that predates the major data center buildout.Lower Umatilla Basin Nitrate ContaminationOregon Groundwater Management AreasU.S. Department of EnergyThe federal record documents BPA's Morrow Flat Substation construction and related transmission improvements near Boardman.Morrow Flat Substation Construction and Related ImprovementsPort of MorrowPort and DEQ records are helping me trace the wastewater side of this story, including data center cooling water wastewater, the Port's industrial wastewater system and land application on agricultural land.Umatilla Electric Cooperative (UEC)UEC information is part of figuring out the local electrical system, how individual data centers were connected and how new infrastructure costs were divided.There are still records I am waiting on, especially the ones that could answer some of the harder questions about who paid for infrastructure, how individual facilities were connected, and how much water is actually moving through these systems.As those come in, I'll add them here and update the findings when they change what we know.

About this case study

This is an ongoing investigation, and I am learning things as I go.I try to keep a pretty clear line between what the records actually show, what looks interesting but still needs more digging, and what I just do not know yet.That matters a lot in Boardman because there are so many things happening in the same place. Agriculture, food processing, data centers, groundwater problems, wastewater, electrical infrastructure and new development all overlap here. Proximity is interesting, but it is not proof that one thing caused another.When new records come in, I will update this page. And if something I find changes the way I understand an earlier piece of the story, I will update that too.Where I can, I will link the original records so you can look at them yourself.Last updated: August 27, 2026Clouds on the Range is an independent research project looking at how data center growth intersects with agriculture, water, energy, infrastructure and rural communities.

← Back to Clouds on the Range