Prineville and Boardman

The corridor that the hyperscalers found first

Map: The Oregon data-center corridor: Prineville and Boardman

The Columbia Plateau unfolds east of the Cascades in long horizontal bands: rimrock, sagebrush, pivot-irrigated fields, and river bottomland. The towns it holds are small. Prineville, the seat of Crook County, had roughly 10,000 residents when Meta began grading its first server farm on the east side of town in 2010. Boardman, in Morrow County on the south shore of the Columbia River, had under 4,000. Both towns were, by any measure, marginal to the national economy.¹

Cheap power changed that. The Bonneville Power Administration, the federal agency that markets electricity from the Columbia and Snake River hydropower system, sells at wholesale to preference utilities across the Pacific Northwest at rates that remain among the lowest in the country. Pacific Power serves Crook County. Umatilla Electric Cooperative serves Morrow County. Both buy BPA power at wholesale and pass the advantage to industrial customers who sign long-term load agreements. A hyperscale data center, drawing 100 to 300 megawatts around the clock, is precisely the customer a small utility welcomes, because the fixed infrastructure costs spread across a predictably enormous load. The math was the same in both counties, and the decisions that followed the math were made in corporate real-estate offices far from the Columbia Plateau.

Meta chose Prineville in 2009. The company broke ground in 2010, placed its first servers in late 2011, and never left. Apple announced its Prineville campus in 2012 and brought it online in 2014. Amazon Web Services began building in Boardman in 2011 and expanded to thirteen separate facilities across roughly 1,000 acres of converted farmland. Business Insider’s 2025 national data-center inventory confirmed that Crook and Morrow counties rank among the earliest non-urban hyperscale anchors in the United States.¹ What the Business Insider map cannot show is what the arrival cost.

This essay treats Prineville and Boardman as one corridor because the mechanisms of extraction connect them. Both counties used Oregon’s enterprise-zone and Strategic Investment Program tax waivers to attract the campuses. Both waived property taxes for fifteen years on capital investments above the program thresholds, directing the cost onto school districts that lack a comparable industrial base of their own. Both drew on water systems where tribal treaty rights and downstream agricultural users held senior claims. The operators are different. The communities are different: Crook County is predominantly rural and white; Morrow County’s Boardman community is predominantly Mexican-American, the descendant of a farm-labor population that arrived under the Bracero Program in the 1940s. The organizing coalitions that formed in response to each campus are different. The injury, however, runs through a single policy architecture, and the present-day fight to reform that architecture is one shared campaign.

What Prineville held and what Meta and Apple took

Map: Prineville and the Crooked River sub-basin before Meta

Crook County takes its name from General George Crook, who campaigned against the Northern Paiute and the Nez Perce across this plateau in the 1860s and 1870s. The Confederated Tribes of Warm Springs, whose reservation lies thirty miles to the northwest, ceded the Deschutes River basin under the 1855 Treaty of Dayton. The treaty’s fishing and gathering protections survived the cession; the Warm Springs Natural Resources Branch administers the tribe’s instream flow and first-food interests on the Deschutes system today.²

Prineville grew as a ranching and timber town. The Crooked River, which drains south from the Ochoco Mountains and joins the Deschutes at Lake Billy Chinook, supplied the irrigated pasture that the ranching economy required. By the time Meta arrived, the Crooked River already ran at reduced baseflow because of the upstream irrigation withdrawals that agriculture had accumulated over a century. The redband trout that spawn in the cold springs feeding the river lower reaches depended on the remaining baseflow to maintain the temperature and dissolved oxygen conditions their eggs require.³

Meta acquired approximately 450 acres of former ranch and industrial land on the east edge of Prineville in 2009. The campus required large volumes of cooling water, and Prineville’s municipal supply is the only practical source for that scale of industrial use in Crook County. The city and Meta negotiated an aquifer storage and recovery agreement: Meta and Apple, which eventually built an adjacent 360-acre campus, fund a system that injects treated surface water into the Crooked River sub-basin aquifer during the high-flow winter months and withdraws it for campus cooling during the summer. The campus operators present the arrangement as conservation, because it avoids direct year-round surface withdrawal.³

Biologists who reviewed the arrangement for the Central Oregon Informed Angler concluded otherwise. Aquifer storage and recovery operations that inject at scale and withdraw in summer reduce the natural baseflow that the river gains from groundwater during the dry season, when snowmelt has finished and precipitation has not yet resumed. The springs that supply the critical cold-water refugia for redband trout lose pressure when the aquifer’s summer water table falls. The biologists estimated that the combined Meta and Apple aquifer storage and recovery operations altered the baseflow regime of the Crooked River at a magnitude large enough to affect spawning conditions in the lower reaches.³ The Oregon Water Resources Department processed the water-right transfers without a separate instream flow impact analysis specific to the Crooked River springs.

WaterWatch of Oregon, which monitors Oregon water-right applications as a public-interest intervenor, documented the Crooked River baseflow reductions in a broader report on Oregon’s data-center water practices. The report named the Warm Springs treaty interest on the Deschutes system as a downstream concern: if the Deschutes loses tributary contributions from the Crooked River because the Crooked’s sub-basin aquifer has been drawn down for data-center cooling, the instream flow at Warm Springs usual-and-accustomed fishing sites falls below treaty-protected levels.² The Warm Springs Natural Resources Branch has not yet filed a formal protest at the Oregon Water Resources Department, but the legal standing for one exists.

The tax mechanism extracted more than water. Oregon’s Strategic Investment Program suspends property taxation for fifteen years on capital investments above 500 million dollars, with a payment in lieu of taxes that counties negotiate individually. Meta and Apple both qualified for SIP treatment. For the first decade of operation, neither paid property taxes at rates comparable to what an equivalent assessed value generated under the standard schedule. Crook County School District, which serves roughly 4,400 students in a county with no other large industrial employer, received the payment-in-lieu amount rather than the standard levy.

In 2024, the accumulated abatement cost the Crook County school district more than sixteen million dollars, according to Central Oregon Daily’s review of Oregon Department of Revenue records. The figure represents the difference between what Meta and Apple paid under the SIP negotiated terms and what the assessed value of their combined campus would have generated at the standard mill rate.⁴ Crook County’s total assessed commercial and industrial value outside the SIP properties is not large enough to replace that loss. The school district has operated on a constrained budget for the duration of the abatement window, cutting programs and deferring maintenance during years in which Meta and Apple reported combined revenues in the hundreds of billions of dollars. A statewide study published by the Oregon Capital Chronicle in August 2025 found that enterprise-zone and SIP abatements across Oregon cost school districts 275 million dollars in a single year; the Crook County loss was a disproportionate share of that total for a county of 25,000 people.⁵

The inversion was specific and mechanical. The city sold water to the campuses at rates designed to attract industrial customers, rates that subsidize industrial use above what the cost of service warrants when the residential base bears the fixed infrastructure costs. The state suspended the property taxes that would have paid for the schools whose students lived in the same houses connected to the subsidized water system. The campus operators kept the margin.

What Boardman held and what Amazon took

Map: Boardman farmland and the Lower Umatilla Basin before Amazon

Boardman sits at the Columbia River’s bend where the river turns west from its north-flowing course through the Hanford Reach. The Port of Morrow controls most of the industrial and agricultural land between the town and the river. The port was created by the Oregon Legislature in 1907 to promote Columbia River commerce, and it has administered irrigation water, industrial sites, and wastewater disposal for the surrounding county ever since.

The Confederated Tribes of the Umatilla Indian Reservation hold treaty fishing rights at usual-and-accustomed sites on the Columbia downstream of Boardman. The Umatilla Reservation, established under the 1855 Treaty of Walla Walla, lies forty miles to the east. The treaty reserved the right to take fish at usual and accustomed places in common with citizens of the United States. The Columbia River sites downstream of Boardman are among the most productive remaining salmon and steelhead habitat in the lower river system. The Tribes’ Natural Resources Department monitors water quality at those sites because nitrate loading from the Umatilla Basin aquifer reaches the Columbia through baseflow and tributary discharge.⁶

The Mexican-American population of Boardman and the surrounding county descends from laborers who arrived under the Bracero Program, the United States-Mexico guest-worker agreement that ran from 1942 to 1964 and that brought farmworkers to the Columbia Plateau’s potato, onion, and asparagus fields. When the Bracero Program ended, many families stayed, built communities in Boardman and Hermiston, and continued working in agriculture. By the 2020 census, Boardman was 43 percent Hispanic or Latino; Hermiston, the larger city twenty miles east, was 39 percent. The farmworker population that formed around the Port of Morrow’s irrigated acreage was the domestic descendant of a labor import program, settled into a county whose groundwater they drew from private wells because the public water system did not reach every rural address.⁶

Amazon Web Services began acquiring parcels through the Port of Morrow in 2011. The port sold former circle-pivot potato and onion ground at prices that reflected agricultural land values, not hyperscale data-center development potential. Amazon built thirteen facilities across roughly 1,000 acres and drew cooling water from the Port of Morrow’s industrial water system. The port had long used a spray-irrigation method to dispose of wastewater from food-processing operations: treated effluent, which carries elevated nitrate concentrations from agricultural processing, is sprayed onto fields adjacent to the aquifer recharge zone. As Amazon’s cooling operations grew, the cooling towers concentrated the nitrates already present in the port’s wastewater stream, and the port continued to spray the concentrated effluent onto the recharge fields. The nitrates percolated through the soil into the Lower Umatilla Basin aquifer.⁷

The Lower Umatilla Basin Groundwater Management Area had been declared by the Oregon Water Resources Department in 1990, after agricultural nitrate contamination first brought the basin above the federal maximum contaminant level of ten milligrams per liter. The GWMA designation required monitoring and directed state management responses when concentrations rose. It did not constrain the port’s spray-irrigation volumes, and it did not account for the industrial-scale cooling tower operations that began two decades after the declaration. Concentrations in the worst-affected wells reached eighty milligrams per liter, eight times the federal limit.⁶

High nitrate concentrations in drinking water cause methemoglobinemia, a condition in which hemoglobin binds nitrate rather than oxygen, in infants under six months of age. The condition is potentially fatal. Adults with compromised kidney function face elevated risk. The Centers for Disease Control links chronic exposure to increased rates of certain cancers. For the farmworker families whose private wells drew from the contaminated aquifer, there was no alternative supply: the public water system did not reach their rural addresses, a filter adequate to remove nitrates costs several hundred dollars and requires regular replacement of media, and the county had not offered subsidized testing or treatment.⁷

Rolling Stone reporter Kiera Butler documented the situation in a feature published in early 2026, drawing on Oregon Rural Action’s well-testing records and interviews with Morrow County residents. The story named the mechanism: Amazon’s cooling towers did not directly contaminate the aquifer, but they amplified the nitrate load in the port’s wastewater stream, and the port’s spray-irrigation recharge practice moved that load into the groundwater that the surrounding community drank.⁶ Oregon Rural Action and Columbia Riverkeeper filed the class-action complaint on behalf of approximately 45,000 residents whose private wells drew from the contaminated zone.

Who fought back and what survived

Map: Organizing response across both counties

The organizing response to the Prineville campuses developed more slowly than the response to Boardman, in part because the mechanisms were less visible. Water-rate inversion and tax abatement produce no air permit, no permit hearing, and no publicly visible construction drama. The damage is distributed across a school budget and a utility rate schedule that most residents read as background facts. WaterWatch of Oregon and the Central Oregon Informed Angler carried the early documentation, because their members are anglers and river advocates who noticed the Crooked River baseflow change before the financial mechanism produced a public fight.

The Central Oregon Informed Angler published its assessment of the Prineville water draws in December 2021, naming the aquifer storage and recovery mechanism and describing the biologists’ findings on redband trout habitat in the Crooked River springs. The assessment reached a local audience of trout anglers and irrigation-water users who understood the river’s hydrology and who did not accept the campus operators’ framing of the aquifer storage system as neutral or conservation-positive.³ WaterWatch of Oregon extended the analysis in its statewide data-center water report and raised the Warm Springs treaty interest as a dimension the state had not addressed.²

The school-tax fight developed a different constituency: parents, teachers, and county commissioners who had watched the school district cut programs while the campus operators reported record revenues. The Crook County school board passed resolutions requesting SIP reform in 2023 and again in 2024. The board’s public record, reviewed by Central Oregon Daily, documented that the district had deferred school building maintenance across a decade that corresponded almost exactly with the SIP abatement window.⁴

In Morrow County, the Boardman organizing followed a model that the atlas has seen before. The community most directly injured, the Mexican-American farmworker families whose wells were contaminated, lacked the institutional infrastructure to self-organize against a defendant of Amazon’s scale. Oregon Rural Action, founded in 2002 in La Grande as a rural-chapter organizing network, filled that gap. Staff organizers drove county roads identifying households with private wells and persuading residents to participate in testing. The door-to-door method is labor-intensive and slow, but it produced a dataset of documented contamination cases distributed across the aquifer’s recharge zone, which is precisely the evidence a class-action complaint requires.⁸

Columbia Riverkeeper joined as co-counsel, bringing its Columbia Basin litigation experience to a case that connected the well contamination to the treaty rights of the Confederated Tribes of the Umatilla Indian Reservation downstream. The Tribes’ Natural Resources Department contributed technical analysis of the contamination pathway from the basin aquifer through baseflow to the Columbia River fishing sites.⁸ The Umatilla tribal interest transformed the case from a local well-water dispute into a treaty-rights matter, which raised the political and legal stakes for Amazon considerably.

Amazon settled in March 2026 for 20.5 million dollars, covering approximately 45,000 residential well owners in the affected zone. The settlement funds in-home well testing, point-of-use nitrate filter installation, and a remediation reserve for future well-owner claims.⁶ The Oregon Capital Chronicle reported that the settlement did not require Amazon to change its cooling tower operations or the port’s spray-irrigation practice; the parties left the ongoing contamination source outside the agreement’s terms.⁸ The class received money. The aquifer remained under the same management regime that produced the contamination.

The Confederated Tribes of Warm Springs Natural Resources Branch has not filed a formal instream flow protest at the Oregon Water Resources Department as of April 2026, but the branch has engaged in the statewide policy conversation about data-center water use that WaterWatch of Oregon has sustained. The Tribes hold senior water rights on the Deschutes dating to the 1855 treaty; those rights predate every agricultural and industrial right in the basin and carry priority in any shortage proceeding. Whether the Warm Springs branch exercises those rights in a formal proceeding against the Crooked River aquifer storage and recovery operations will depend on whether the Deschutes system falls below the flow levels the treaty contemplated, a threshold that the branch monitors annually.²

Present-day parallels and the shared organizing front

Map: Live fights across the Prineville-Boardman corridor

The Prineville and Boardman cases share a policy architecture, and the present-day fight against that architecture is one campaign conducted on two fronts. The Oregon Legislature carries the SIP reform fight. Oregon Rural Action and Columbia Riverkeeper carry the aquifer fight. WaterWatch of Oregon carries the instream flow fight. The Confederated Tribes of the Umatilla Indian Reservation and the Confederated Tribes of Warm Springs provide the treaty standing that elevates each fight to a federal dimension.

The Strategic Investment Program reform fight intensified in the 2024 legislative session, after the Oregon Capital Chronicle published the 275 million dollar statewide cost figure and Central Oregon Daily published the Crook County school loss in the same year.⁵⁴ Bills proposing to shorten the abatement window, to require community benefit agreements as a condition of SIP approval, and to direct a portion of the payment-in-lieu to school districts at rates closer to the standard mill calculation all reached committee hearings. None passed in 2024. The 2025 session returned to the same material, with the Oregon School Boards Association joining the coalition that had previously consisted primarily of rural county commissioners and water advocates. The reform campaign continues in the 2026 session.

The senior water-rights strategy at the Oregon Water Resources Department offers a different lever. Oregon’s water law is prior appropriation: the oldest right governs in shortage. The Confederated Tribes of Warm Springs hold 1855 treaty rights that predate every agricultural and industrial water right in the Deschutes and Crooked River basins. If WaterWatch of Oregon documents that the Meta and Apple aquifer storage and recovery operations reduce Crooked River baseflow below the levels the treaty contemplated at Warm Springs usual-and-accustomed fishing sites, the Tribes could petition the Oregon Water Resources Department for a curtailment order against the junior industrial rights. The petition mechanism exists, has been used against agricultural users in other Oregon basins during drought years, and would function identically against a data-center operator.²

The settlement implementation fight in Morrow County is an immediate organizing target. The 20.5 million dollar Amazon settlement fund requires a claims administrator, a testing and installation program, and ongoing monitoring of the affected wells. Oregon Rural Action has the community relationships and the door-to-door testing infrastructure to participate in settlement implementation in a way that ensures the money reaches the farmworker families with the most severe contamination. The risk, which the organization has named in its public communications, is that the claims process will favor English-speaking property owners with access to legal assistance and that Spanish-speaking renters and informal-title holders in the most contaminated zones will not navigate the claims process without organizational support.⁷

The atlas has documented this risk before. The Cabrillo Village cooperative purchase in Ventura County is the model case in this collection for a Mexican-American farmworker community that organized through a contaminated-water framework and forced a corporate defendant to terms. The Cabrillo Village residents won not because they had money or political connections but because they built a community organization capable of sustained participation in a legal process that a corporate defendant could outlast if the community had no institutional anchor. Oregon Rural Action plays the same institutional role in Morrow County that the United Farm Workers organizing infrastructure played in Ventura County.

The Chávez Ravine case in the atlas provides a second parallel, applicable to Prineville rather than Boardman. Chávez Ravine documents how the City of Los Angeles traded a public resource (publicly owned land cleared of a Mexican-American community under the Housing Act of 1949) to a private operator (Walter O’Malley and the Dodgers) on terms the public could not review until decades after the transaction. The Prineville water-rate inversion and SIP abatement operate through the same structural logic: a municipality and a state agency trade a public resource (municipal water at below-cost rates, school tax base through abatement) to a private operator on terms the public records disclose only through investigative journalism and public-records requests. The mechanism is quieter than bulldozers. The transfer is equally real.

The San Pedro essay in this atlas describes how the Terminal Island fishing community built a documentary record of its displacement that the state eventually had to confront in the Channel Islands renegotiation. The Prineville and Boardman documentary record is now denser than that record was at the equivalent moment after displacement. Central Oregon Informed Angler published the biological findings in 2021. WaterWatch of Oregon published the statewide policy analysis in 2025. The Oregon Capital Chronicle published the school-cost figures in 2024 and 2025. Oregon Rural Action holds the well-testing database. Columbia Riverkeeper holds the litigation record. The Food and Environment Reporting Network published the agricultural-context analysis. Rolling Stone published the community narrative that made the Morrow County story visible to a national audience.⁶

What the corridor does not yet have is a single institution whose job is to maintain the archive and make it available to future organizers and future legal proceedings. The Boxtown case in this atlas shows what that institution looks like when it exists: Memphis Community Against Pollution maintains the public-records request files, the air-quality monitoring data, and the permit-challenge submissions in a publicly accessible archive at memphiscap.org. The Black Bottom Archives in Detroit maintains the oral-history and property-record collection that now anchors the I-375 removal negotiation. For the Prineville and Boardman corridor, the existing organizations together hold the records; no single institution yet serves as the archive for the corridor as a whole. Oregon Rural Action has the community relationships to play that role in Morrow County. The Central Oregon Informed Angler has the biological record for Crook County. The 2026 organizing calendar offers the settlement implementation process as the organizing moment at which those records can be consolidated and made permanent.

The corridor’s two communities share one further attribute that this atlas treats as an organizing asset: they were found, not assembled. The Mexican-American farmworker families of Boardman did not arrive because Amazon was coming; they arrived sixty years before Amazon, under a labor-import program that deposited them in a county without the political power to protect its own groundwater. The ranching and timber families of Crook County did not organize to attract Meta; Meta chose Prineville because cheap power and an accommodating city government made the site available. In both cases, the community’s prior presence is the moral ground on which the present fight stands. The hyperscale operators arrived after the communities existed. The communities have priority, in every sense the word carries in Oregon water law and in the ordinary sense of who was there first.

Footnotes

1. Business Insider, “See where data centers are across the US on our interactive map,” September 2025, https://www.businessinsider.com/data-center-locations-us-map-ai-boom-2025-9. Cited for the confirmation that Crook and Morrow counties rank among the earliest non-urban hyperscale data-center anchors in the United States and for the continued scale of the corridor in the 2025 national inventory. ↩ ↩²

2. WaterWatch of Oregon, “How Oregon’s Data Center Boom is Supercharging a Water Crisis,” 2025, https://waterwatch.org/how-oregons-data-center-boom-is-supercharging-a-water-crisis/. Cited for the Crooked River baseflow analysis, the Warm Springs treaty interest on the Deschutes system, and WaterWatch’s intervention record at the Oregon Water Resources Department on data-center water-right transfers. ↩ ↩² ↩³ ↩⁴ ↩⁵

3. Central Oregon Informed Angler, “Prineville, Data Centers, and Water: There is a Cost,” December 2021, https://coinformedangler.org/2021/12/26/prineville-data-centers-and-water-there-is-a-cost/. Cited for the biologist assessment of aquifer storage and recovery operations, the Crooked River baseflow reduction mechanism, and the redband trout spring habitat harms. ↩ ↩² ↩³ ↩⁴

4. Central Oregon Daily, “Corporate tax breaks cost Crook County schools millions,” 2024, https://www.centraloregondaily.com/news/prineville/crook-co-schools-lose-tens-of-millions-of-dollars-to-corporate-tax-breaks/article_3c91119c-91e0-4186-95ee-2c66986d5912.html. Cited for the 16 million dollar 2024 school-district loss, the decade of deferred maintenance, and the school board resolutions requesting SIP reform. ↩ ↩² ↩³

5. Oregon Capital Chronicle, “Oregon business property tax breaks cost schools 275 million last year, study finds,” August 2025, https://oregoncapitalchronicle.com/2025/08/21/oregon-property-tax-breaks-for-corporations-cost-schools-275-million-last-year-study-finds/. Cited for the statewide 275 million dollar school-cost figure from enterprise-zone and SIP abatements. ↩ ↩²

6. Rolling Stone, “Amazon to Pay 20.5 Million in Settlement of Class Action Suit Over Pollution in Eastern Oregon,” March 2026, https://www.rollingstone.com/culture/culture-news/amazon-data-center-pollution-settlement-1235539152/. Cited for the March 2026 settlement amount, the 45,000-resident class covering private well owners in Morrow County, the nitrate concentration reaching eight times the federal limit, the Confederated Tribes of the Umatilla Indian Reservation downstream treaty interest, and the Bracero Program origin of the Mexican-American farmworker population. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

7. Food and Environment Reporting Network, “Amazon to Pay 20.5 Million to Settle Suit Over Pollution in Oregon,” March 2026, https://thefern.org/2026/03/breaking-amazon-to-pay-20-5-million-to-settle-class-action-suit-over-pollution-ineastern-oregon/. Cited for the spray-irrigation recharge mechanism, the nitrate concentration pathway, the methemoglobinemia health risk, the absence of public system access for rural well holders, and the claims-process equity concern for Spanish-speaking residents. ↩ ↩² ↩³

8. Oregon Capital Chronicle, “Amazon to pay 20.5 million settlement over northeast Oregon nitrate pollution,” March 2026, https://oregoncapitalchronicle.com/2026/03/31/amazon-to-pay-20-5-million-settlement-over-northeast-oregon-nitrate-pollution/. Cited for the settlement terms, the absence of operational-change requirements, and Oregon Rural Action and Columbia Riverkeeper’s organizing and litigation roles. ↩ ↩² ↩³