Data centers are buildings full of computers that store information and run online services. They bring jobs and tax payments, but they also need electricity and water. Oregon offers some of these projects tax breaks. To judge whether a break is worth its cost, we need to ask: would the company build here without it?
The agreement sets what the company pays.
Public costs also depend on its electricity, water and service needs.
A company needs more than a tax agreement to build a data center. It also needs permission to use the site and a way to get electricity. Separate public decisions govern each step.
On June 5, Oregon paused new data-center approvals under its standard enterprise-zone program, which temporarily reduces property taxes. Two other tax-break programs remain available: long-term rural enterprise zones and the Strategic Investment Program.
Business Oregon · program comparison ↗Portland General Electric (PGE) now charges data centers under a separate set of electricity rates. Other power companies have their own rules, so customers elsewhere may have different protections against paying for data-center costs.
PGE · July 8 rate changes ↗On July 27, Hillsboro paused new applications for sites whose main use would be a data center or battery storage. Applications submitted before the pause may still move forward.
Hillsboro · 120-day pause ↗House Bill 4084 pauses new data-center approvals under the standard enterprise-zone program until 90 days after lawmakers finish their 2027 session. Projects approved earlier fall into a separate category. Business Oregon, the state’s economic development agency, still lists data centers as eligible for the long-term rural program and the Strategic Investment Program, often called SIP.
The POWER Act sets requirements for electricity companies owned by investors. The Oregon Public Utility Commission, the state regulator, oversees their rates and service rules. PacifiCorp and Idaho Power have separate proceedings to put the law into practice. Utilities owned by customers or local governments follow different rules. A June schedule expected a PacifiCorp decision in October and a filing showing compliance in November. Those were planned dates; check the regulator’s case record for later decisions. Oregon PUC ↗
A lawsuit filed in June challenges tax-break approvals made in Hillsboro and Washington County before the pause. The people and groups suing say officials did not follow the required approval process. Hillsboro says its staff followed existing program rules. We have not verified a later court ruling that settles the dispute. Reporting on the complaint ↗ The city’s explanation ↗
A tax break can bring in public money if it attracts a project that would otherwise go elsewhere. If the company would build here anyway, the same break lets it pay less for an investment Oregon would already receive.
The taxes and fees could leave the community better off than an empty site would. Those payments still need to cover the added cost of public services.
The company could have made the same investment and paid the full property tax bill. Officials need to explain what the public gains by reducing that bill.
The size of a proposed project cannot tell us whether a tax break is necessary. We need evidence about the company’s other possible locations, their costs, and why it would choose one over another.
of the property taxes collected in Morrow County came from data centers, according to ECONorthwest’s July 2026 analysis.
A September study estimates that Oregon’s recent data-center building activity supported about 3,500 construction jobs per year. That work ends at each project, but workers may move to the next one. Suppliers and service businesses benefit too. See the job counts and what they include.
The Dalles negotiated payments based on a share of what the company would owe without a tax break. Its agreement also sets a minimum annual payment. You can try those rules in the calculator below.
The case is stronger when a company identifies realistic competing sites and shows how the tax break affects its choice. Officials also need checked estimates of the money local governments would receive and spend. An empty site today could still attract another use later.
The $22.96 million goes to several local governments and public districts, rather than all going to the county government. Job estimates should explain how long the work lasts, who gets hired, what workers earn, and whether that work depends on the tax break. ECONorthwest ↗
Oregon uses both local taxes and state money to fund schools. Its funding formula is a set of rules that divides this money among districts, taking student numbers and needs into account. When a tax break reduces a district’s local revenue, the state generally provides more of its funding. If the state’s school budget stays the same, that leaves less money to share across Oregon’s schools.
This is the usual pattern when the state does not add money to its school budget. Some districts and taxes follow different rules.
Oregon Department of Education · school funding ↗Payments that companies make to schools under enterprise-zone agreements count as local revenue in the state’s calculation. They can therefore reduce the amount the state provides.
Some districts already collect more locally than the formula would give them. The state does not replace their lost revenue unless the loss brings them below that amount.
Other rules apply to taxes that repay money borrowed for school construction and to some extra taxes approved by local voters, called local-option taxes. To understand a particular deal, identify which school taxes it reduces and which rules apply to them.
Identify where the water comes from and how much the facility takes.
Measure how much water is used up, how much is returned, and whether its quality changes.
Check the highest summer use and how it would be limited during a drought.
A facility may use a small share of a city’s water over a year but still put pressure on supplies during a dry summer. Officials also need to examine electricity supply, pollution from backup generators, noise, and the combined water use of facilities drawing from the same area. The state’s advisory committee is still examining these questions. Data Center Advisory Committee ↗
A deal is easier to justify when officials can show that the tax break is needed, payments cover public costs, and water and pollution limits can be enforced. Even a large tax break does not tell us how much revenue Oregon could have collected without it: the company might have built elsewhere.
A 2016 state audit divided the value of 2015 tax breaks by the number of new jobs businesses reported. The result was $4,200 per job in standard enterprise zones and $54,500 in long-term rural zones. The calculation does not tell us how many of those jobs would have existed without the breaks. It also does not tell us what a proposed data center would cost today.
These figures cover one year and use dollars from that time. The calculator combines payments over many years and gives future payments less weight, so its totals are not directly comparable.
Source: February 2023 reporting on the 2016 auditA September 2026 study separates the people running Oregon’s data centers from the work of building them. The two estimates cover different years and types of work.
estimated construction jobs per year
Average during 2023–2025
This estimate includes employees and self-employed business owners. It covers work on buildings and systems such as power and cooling. It excludes installing servers and other computing equipment.
Work ends when a project finishes. New projects and upgrades can keep crews employed.
estimated employees running data centers
Annual average in 2024
This estimate covers employees of companies that own, run or rent space in a data center. It excludes staff supplied by other companies, such as contract security and cleaning workers.
These roles continue while facilities operate, though staffing can change over time.
We do not have that breakdown. These counts can include full- and part-time work. “Full-time” describes weekly hours; “temporary” describes how long a job lasts. A construction worker can work full-time on a temporary project. How government job counts work ↗
These are estimates of work supported by the industry, not counts of new hires or proof that tax breaks created those jobs. The study was prepared by ECONorthwest and the University of Virginia, with funding from the Lemelson Foundation.
The operations estimate uses employer records checked against known data-center locations. It is an average across 2024, not a count taken today. The construction estimate uses a model based on the 2023–2025 building activity. The researchers did not have the projects’ construction payrolls, so that figure is less direct. Neither estimate tells us how many workers live in the host community. Methods · pp. 76–82 and 89–93 ↗
The same study estimates another 7,780 jobs supported by suppliers and worker spending connected to operations, and another 4,844 connected to construction. These include work at other businesses, not just at data centers. They should not be described as permanent on-site jobs. Tables 6.1 and 6.6 · pp. 86 and 93 ↗
Choose an example based on a real agreement or program, then change its assumptions. The calculator compares the money governments might receive with a tax break and without one.
Choose an example to fill in every value. Then change the assumptions to see when a tax break brings in more public money.
These examples show what could happen under stated assumptions. The Dalles examples use a signed agreement, and Hillsboro uses published program rules. They do not show verified amounts collected from those projects. Values we do not know are labeled “Assumption.”
The more likely the project is without a tax break, the harder the break is to justify.
This is your estimate, not a measured chance for any of these sites. For example, 50% means you think building without the break is as likely as not building.
Dollar amounts marked M are millions; K means thousands. For example, 600 in a $M field means $600 million.
The signed agreement requires Project 1 to pay 50% of its full property tax bill, with a minimum annual payment of $3 million. It also requires a one-time $3 million payment. Exhibit A gives an example using $600 million of property value at a 1.10% tax rate. Those are illustration values, not actual assessed values used to set a tax bill.
We keep the illustrated property value and tax rate unchanged, count the one-time payment in year 1, and leave out scheduled payment increases. We choose the remaining values, including the chances of building, years of operation, jobs and costs, as assumptions you can change.
Read the original source ↗Each example assumes a 50% chance of construction without a break and a 100% chance with a deal. It also assumes 30 years of operation and $50,000 a year in land taxes if nothing is built. A 4% discount rate gives future payments less weight than money received today. Jobs, wages and state income-tax rates are examples, too. Public costs and construction tax revenue start at zero because we do not have estimates for them.
Counts money received by governments and public districts in the area. It does not calculate changes in state school aid.
The deal brings in $22.8M more than offering no break. This result uses your assumptions and subtracts the public costs you entered.
Amounts are in millions of today’s dollars. They combine the chances of building and not building, rather than showing a guaranteed payment.
If the chance of building without a break is below 70.2%, offering the deal brings in more money in this calculation. Above that chance, offering no break brings in more. The dot marks the chance you entered.
The shaded area shows the years with a tax break. After that, the model charges full property tax for as long as the site operates.
This calculator compares public money, not every effect of a project. Water use, pollution, cleanup costs and other possible uses of the land need their own review. A zero in a cost field means the cost has not been estimated; it does not mean the project has no such cost.
Both options assume the same facility, start year and years of operation. The property value used to calculate taxes stays the same throughout. Multiplying that value by the tax rate gives the full annual property tax. During the break, the deal follows the payment rule you choose. After the break, it charges full tax.
The calculator subtracts the public costs you enter. Taxes from construction and the one-time deal payment are counted in year 1. Future money receives less weight: each year’s amount is divided by (1 + discount rate) raised to that year’s number. At 4%, for example, $104 received one year from now counts as $100 today.
The calculator combines the possible outcomes using the chances you enter. At a 50% chance of building, it counts half the value if the project is built and half the land taxes if it is not. This is an average of possible outcomes, not a promise of payment. After the site closes, only the assumed land taxes remain.
To find the break-even chance, it subtracts the value of leaving the land undeveloped from the deal’s estimated value. It divides that difference by the extra value of a project paying full tax. This calculation applies when a fully taxed project brings in more than leaving the land undeveloped.
The Local + Oregon view estimates annual state income tax by multiplying permanent jobs, annual wages and the average tax rate entered. It does not account for workers leaving other taxable jobs or for tax credits. The model also leaves out changes in school aid, changes in property tax values, corporate taxes, and fees utilities pay local governments. It does not calculate the effects of another development on the site or environmental damage.
The Dalles examples use the signed 2021 agreement with Design LLC, a Google company. Its Project 1 illustration assumes $600 million of taxable property at a 1.10% tax rate. That would produce a $6.6 million annual bill without the break. A 50% payment would be $3.3 million. Project 2 uses a 60% share, and both projects have a $3 million minimum annual payment. Taxes and community-service fees already included in those payments are counted once. The calculator does not include scheduled payment increases or changes in the property’s tax value. Signed agreement · definitions and Exhibit A ↗
The Hillsboro example shows how the city’s program rules could work, using the highest allowed city fees and school payments. The payment share rises after year 3. This is not a forecast for a particular project or a claim that a new project qualifies today.
Morrow County’s 2023 approval record describes a 15-year tax break under the Strategic Investment Program. The deal initially leaves $100 million of property value subject to tax and includes a community-service fee of up to $2.5 million, plus other obligations. A single annual payment would not capture those rules. We need the full payment and property-value schedules before making a project forecast. Approval minutes · page 6 ↗
These are questions we suggest asking in each region. The map shows approximate locations, not exact project sites. It does not show whether a site has enough power, water or internet access.
Does a site with existing internet connections and customers still need a tax break?
Compare similar projects built without tax breaks. Examine the application, available electricity and other industrial uses of the site. Good internet connections may help attract a project, but they do not tell us how much tax relief it needs.
City of HillsboroWhere would an expansion get its water, and how would its payments be calculated?
Check water supply and demand by season, then read the signed agreement. The 2021 Strategic Investment Program agreement sets minimum payments and shares of what the full tax bill would be. A single flat fee would not describe it accurately.
The Dalles / Wasco CountyWould the project’s public payments cover its costs, and who would pay for its electricity needs?
Review the project’s contract, plans to buy electricity and water-quality limits. Compare those terms with realistic alternative locations and uses of the land.
Morrow County AssessorDoes an expansion need the same tax break as the original facility?
Ask for evidence about the expansion’s costs and possible locations. Check the legal right to use water and the power company’s current rates and service rules.
Data Center Advisory CommitteeWhat else could use this land, water supply and electricity connection?
Compare industrial, farming and other uses allowed on the particular site. A state decision about one property does not settle whether data centers belong elsewhere in the valley.
Governor's OfficeCan electricity, high-speed internet and water actually reach the proposed site?
Ask for utility studies, connection costs, access to fiber-optic internet lines and evidence of local water supplies. These are needed before concluding that the site can support a data center.
Data Center Advisory CommitteeA project may create sales, wages and construction work while bringing in less tax revenue than the tax break costs. A study of those benefits also needs to ask how much would have happened without the break.
The charts compare three programs. “Standard” means standard enterprise zones; “Rural” means long-term rural enterprise zones; and “SIP” means the Strategic Investment Program. All three offer property-tax breaks under different rules.
The study adjusts the cost of the tax breaks to account for required fees and other included payments that governments still receive. The charts use that adjusted cost in both comparisons.
The study estimated the value of goods and services produced by activity linked to each tax-break program. It compared that value with the program’s adjusted tax-break cost.
The second comparison counts only state income tax from employees. Zero means those taxes equal the adjusted tax-break cost. A negative number means they fall short.
0 = taxes equal the adjusted cost
These historical figures cover whole tax-break programs, including businesses other than data centers. February 2022 study · figures 17 and 20 ↗
The study calls these figures “modified return on investment,” or ROI. It compares an estimated benefit with the cost of tax breaks after subtracting required fees, application fees and included taxes on property outside the exemption. It divides the benefit by that adjusted cost, then subtracts one. A result of zero means the two amounts are equal; a positive result means the measured benefit is larger.
One comparison measures the value of goods and services produced. The other counts only state income taxes from employees. Neither is a full account of public benefits and costs: several business taxes and environmental costs are left out.
For the rural program, the study’s income-tax calculation is $81.46 million ÷ ($534.95 million − $34.20 million) − 1, or about −0.84. That means about 16 cents in employee income tax for each dollar of adjusted tax-break cost. The standard program’s +1.35 result means $2.35 in employee income tax for each dollar of adjusted cost. These estimates do not tell us how much investment the tax break caused. Business Oregon repeated the business-activity figures in June 2026; the income-tax figures come from the original study.
Of the $15.8 billion invested through the long-term rural enterprise-zone program, $15.4 billion went to data centers. Other programs also support data centers, so this is only part of Oregon’s total.
Business Oregon · slide 4 ↗The same table lists another $3.1 billion in data-center investment under standard enterprise zones and $11.2 billion under the Strategic Investment Program. The $15.8 billion total shown above belongs only to the rural program.
Business Oregon reports 7,600 “direct” jobs for 2025. The September study counts 2,629 employees linked to operating facilities in 2024. We have not established exactly which workers explain the difference.
The larger figure cannot simply be labeled permanent jobs, and the gap cannot be assumed to represent construction workers. Check the year, employers and types of work each source includes.
Business Oregon · slide 2 ↗ September study · definitions, pp. 74–82 ↗PGE’s new rates took effect July 8, 2026, under the POWER Act, Oregon’s law addressing large electricity users.
The new rates raised this group’s average electricity rate by 29.7% on July 8, 2026. The rules also use contracts to help keep data-center costs from falling on other PGE customers.
This change lowered PGE’s average residential rate by 1.3%. A household’s bill still depends on how much electricity it uses and any other rate changes. These figures do not apply to every Oregon utility.
PGE’s new rates shift more electricity costs to data centers. The result does not tell us how other utilities divide their costs or whether the higher rates will affect future projects.
The records we reviewed show some protections for the first two tests, but they do not establish that any test is met across Oregon. Each proposed project needs to show how it would meet these standards.
Require the project to cover the electricity costs it creates, including costs left behind if it closes early.
PGE has introduced separate data-center rates and contracts intended to protect other customers. Other power companies have different rules and agreements that need their own review. Read the source ↗
Identify where the extra electricity will come from and when it will be available. Explain how the supply will meet clean-energy requirements.
PGE checks whether a new connection can meet Oregon’s clean-electricity requirements under House Bill 2021. That check does not show that clean power will be available every hour or that other utilities face the same requirements. Read the source ↗
Set measurable limits on water use, pollution and noise. Explain what happens during a drought or periods of highest demand.
The committee still has questions about water use and the combined effects of several facilities in the same area. One company’s claim that it uses water efficiently cannot answer those questions for the whole state. Read the source ↗
Compare public payments and costs with and without the tax break. Include realistic alternative uses of the land.
Existing studies do not tell us how likely each project is to be built without a tax break. They also do not count every public payment and cost needed to judge a project’s financial benefit. Read the source ↗
Explain how the deal affects local and statewide school funding, and how added public services will be paid for.
Oregon’s school-funding rules can spread the cost of a local tax break across the state. Some voter-approved taxes, school construction debt and districts with unusually high local revenue follow different rules. Read the source ↗
Publish the agreement and the evidence supporting it. Identify who will check each promise and what happens if it is broken.
The Hillsboro lawsuit alleges problems with particular approvals. It does not establish that all Oregon tax-break approvals were unlawful. Each program has its own approval rules. Read the source ↗
These are standards we propose for future decisions. They are not all required by current law, and an individual project may already meet some of them. A project must still follow water laws and approval rules even if it brings in more tax revenue.
Oregon’s Data Center Advisory Committee advises the governor on data-center policy. Its September report sets out early findings and questions for the public, with final recommendations expected before the end of 2026.
The September report explains what the committee has learned and what it still wants to know.
September preliminary findings ↗Explain what you observed and how you made any estimates. Include a source and say what action you would like officials to take.
The official committee page links the written-comment form and any schedule changes.
Committee notice ↗The governor originally asked for recommendations in October. The committee’s September report says it now expects to finish before the end of 2026. Read the updated report.
Check the official notice and written-comment form for current instructions.
The pause on new standard enterprise-zone approvals lasts until 90 days after the 2027 legislative session ends. Changes to other tax-break programs and utility rules require separate decisions. For mailing-list updates, the Oregon Department of Energy lists datacenter.ac@oregon.gov. Hearing from a tribal representative at a public meeting does not replace formal consultation between the state and a tribal government.
The material below includes government estimates, company statements and our own analysis. We identify who made each claim so you can judge its support. Inclusion in a public hearing does not mean a claim has been independently checked.
Business Oregon’s June table reports $233.6 million in data-center tax breaks out of $240.9 million for the long-term rural program. That amount covers one program, not all data-center tax breaks in Oregon.
The Morrow County assessor reports $123.6 million in taxes companies did not have to pay, $72.3 million in taxes collected across the county, and $23.9 million in negotiated fees. These figures measure different things. The tax-break total is not the value of the property, nor does it show how much the county would collect if projects chose to build elsewhere.
ECONorthwest estimates that data-center electricity use could rise from 14.0 terawatt-hours in 2025 to 24.8 in 2030. One terawatt-hour is one billion kilowatt-hours, the unit used on electricity bills. The estimate is preliminary: actual use depends on what gets built, how much it runs, and how efficiently it operates.
The Oregon Department of Environmental Quality lists 2,482 generators at 39 data-center sites. Their permits allow a combined 6,328 megawatts of generating capacity; a megawatt measures how much power equipment can produce at a given time. To estimate annual pollution, we also need their hours of use, fuels, pollution controls and permit limits.
The Department of Education describes different rules for everyday school funding, taxes for construction debt and some extra voter-approved taxes. Districts with enough local revenue to exceed their formula funding also receive different treatment. A project review needs to identify exactly which taxes would be reduced.
The committee is asking for public input before issuing final recommendations. Its September report records questions and areas of agreement, but does not itself create new legal requirements.
State economic development agency
The agency says tax breaks are meant to attract investment and jobs.
Its June presentation shows which programs businesses use and the investment and jobs reported. Those figures do not tell us whether each project would have happened without a tax break. Published source
Published enterprise-zone explanation
The city says staff handled tax-break applications under its existing rules.
This is the city’s account of the approvals; a court must decide the legal challenge. Approval for a tax break is also separate from permission to develop a site. Published source
Umatilla Electric Cooperative
The cooperative says its locally governed contracts can protect existing customers.
The cooperative is owned by its members. It says it has experience keeping data-center costs from falling on other customers. To assess that claim, examine the contracts, who pays each cost, and how future electricity will be supplied. Published source
March water presentation
Amazon Web Services says it is reducing water use for cooling and investing in community water systems.
These benefits are reported by the company. To assess the effect on local supplies, compare water taken and water restored in the same area and season. Restoring water elsewhere or at another time may not address a local shortage. Published source
Tax Fairness Oregon
The organization wants to end property-tax breaks for data centers.
Its testimony questions whether the breaks are needed and who benefits from them. That is a policy position; it does not prove that every project would go ahead without a break. Published source
Confederated Tribes of the Umatilla Indian Reservation
The statement calls for formal consultation between the state and tribal governments.
Tribes are governments with their own authority and rights under treaties. A representative’s participation at a public meeting does not replace formal consultation about recommendations that affect those rights and resources. Published source
Tax breaks, electricity rates, permission to build and environmental permits have separate approval processes. A proposal needs to identify who can set each requirement and who can enforce it.
A community may gain tax payments while effects on school funding, water supplies or the electricity system reach other places. Each of those effects needs to be measured. A local benefit does not settle the question for the rest of Oregon.
A public hearing tells us what its participants said, not what everyone in Oregon thinks. Serving on the committee, submitting a comment and taking part in formal tribal consultation are different ways to participate.
Public opposition and competing proposals show that people disagree about policy. They do not prove that a particular change is certain to pass or that nobody supports a pause on development.
We assembled this reading list in August 2026 and added the committee’s September report. For newer material or updated links, use the official committee index.
A recording is available; no slides were posted.
Portland Civic Lab prepared this analysis from published records. We checked the evidence on September 29, 2026. Older figures are labeled with the dates they describe.
A source reports a decision, number or term.
A study uses data and calculations to estimate a result.
We choose a value where information is missing. You can change it in the calculator.
We explain what we think the evidence means or what officials should require.
We checked selected original records to understand how the tax programs work, what the financial and job figures count, and how the public can comment. The document index helps readers find more material. We have not transcribed every recording or independently checked every statement.
The views we describe come from published material; we did not conduct new interviews. We still need complete schedules of project property values and payments, a breakdown of full- and part-time jobs, an explanation of the differing job estimates, and updates on court and utility decisions. The calculator combines documented rules with labeled assumptions. Its results are examples, not verified amounts that governments have collected.
Send a correction or documented response with the claim, source page and proposed correction.
