Housing Supply - Salem River District

The Gap

What the three states say they need, against what they build

The constraint here is delivery. Every West Coast state now publishes a legal housing production target, and production is running well below each of those targets. These are the states’ own numbers.

State	Target a year	Built in 2024	Short by
California	~312,500	~100,400	~212,000
Washington	~55,000	~37,000	~18,000
Oregon	29,522	~14,000	~15,500
Together	~397,000	~151,000	~245,000

California's figure is the sixth-cycle Regional Housing Needs Allocation of roughly 2.5 million homes across eight years. Washington's is the Commerce Department's 1.1 million over twenty years under HB 1220. Oregon's 29,522 is the official annual target published by the Office of Economic Analysis under the Oregon Housing Needs Analysis; the Governor's separate goal is 36,000. Production figures are Census permit data for 2024. The annual conversions and the total are our arithmetic, not any agency's published number, and California's 2.5 million total has been criticized by the State Auditor as not reproducible. Treat the shape as solid and the decimals as approximate.

The Interstate 5 metros from Seattle to San Diego hold roughly 34 million people. No agency publishes a corridor housing figure, so that is our aggregation of Census metro estimates.

The Record

The last generation raised billions and mostly failed

Any serious version of this argument has to start with the graveyard, because the graveyard is recent and it is large. We would rather put it at the top of our own page than have someone else put it in front of us.

Katerra

Raised about $2 billion, largely from SoftBank, at a $4 billion valuation. Bankrupt in June 2021. It tried to integrate design, manufacturing and general contracting at once, before any one of them worked.

Veev

Raised several hundred million and reached unicorn valuation in 2022. Shut down in November 2023 when a funding round fell through. Its assets went to Lennar in a fire sale.

Entekra

Backed by Louisiana-Pacific with a $35M factory rated for 3,000 house frames a year. Closed in 2023. Reported output stayed below what the plant was rated for.

Factory_OS

California's flagship modular plant at Mare Island, restructured and renamed in 2024. Filed a notice in February 2026 threatening to lay off all 280 employees.

Connect Homes

Closed its only California factory in March 2024 and moved to Arizona, leaving customers mid-order.

S2A Modular

Raised roughly $65 million from around 350 retail investors for factories that mostly went unbuilt. Its founders were sued by the SEC in July 2026.

Each of these was a delivery failure. The waiting list was never the problem. They failed against the five constraints in the next section, and those constraints are structural rather than managerial, which is why more capital made the failures faster rather than smaller.

The Constraints

Five problems, and who can actually fix each

Each constraint has a company answer, already tried, and a state answer that is available to Oregon today. The company answers are why the graveyard exists. Whether the state answers work is the open question, and it is the reason we think a region rather than a firm might be the right unit to attack this with.

Constraint

The company answer, already tried

The state answer, not yet tried here

Factory utilizationThe killer

A plant needs 70 to 80 percent throughput to break even, and construction demand is lumpy, seasonal and rate-sensitive. Every failure above built capacity ahead of committed orders.

Aggregate demand into a standing order book. Public housing agencies, universities, health systems, cities and several developers, pooled into predictable annual volume. This is exactly what the durable European producers have and no American company has ever had.

Transport distanceThe physics

Whole modules travel 250 to 400 miles economically. Transport runs $15 to $20 a mile before escorts, and buildings are cheap relative to their bulk. HUD found the same ceiling in the 1960s.

Ship components rather than boxes, and change the cost per mile. Flat panels load densely on standard trailers with no oversize permits. Autonomous electric freight on the existing rail line attacks the constraint itself rather than working inside it, and no single company can build that.

Construction financeThe quiet one

Sixty to eighty percent of a building's value is completed inside a factory, months before anything appears on site. Lenders release draws against visible site progress, so the manufacturer carries the work in progress.

A standing facility that lets lenders advance against factory work in progress. Credit enhancement or a state-backed off-site progress instrument. A small policy change with a large effect, and it works for every manufacturer in the state rather than one deal at a time.

Code fragmentationAlready solved

A wall engineered for one state's code does not transfer to another's, and each jurisdiction has its own approval. This is the single biggest reason manufactured housing outperforms modular: it has one federal code.

Oregon already has this. We went looking for a state program that markets it to builders and could not find one, and would be glad to be pointed at it. The Prefabricated Structures Program under ORS 455.705 gives one plan approval and a state insignia valid in every Oregon jurisdiction, with reciprocity agreements already in place with other states.

Capital durationThe mismatch

Venture capital wants scale in five to seven years. This business has low margins, high capital cost, a hard geographic ceiling and a payback measured in decades. Every durable success ran on patient balance sheets instead.

Pension capital is the right duration. Oregon holds one of the largest public pension funds in the country. An in-state allocation to housing production infrastructure matches a thirty-year liability to a thirty-year asset, and it is a decision only the state can make.

Sources for the constraint figures: McKinsey's 2019 work on modular construction for the utilization and savings analysis; oversize transport permit and escort rates from commercial haulers; HUD's Operation Breakthrough evaluation for the historic shipping radius; the Construction Financial Management Association on off-site draw schedules; Oregon Building Codes Division for the Prefabricated Structures Program.

The Position

What Oregon already owns

The unusual thing about this argument is how little of it requires anything new. Four of the five instruments already exist in Oregon. Pointing them at the same target is the move this page describes.

One code

The Prefabricated Structures Program gives statewide plan approval and an Oregon insignia, plus reciprocity with other states. On these two points Oregon is better ground for offsite manufacturing than California is today. We went looking for a state or industry source that makes this comparison and could not find one, so treat it as our reading of the two statutes.

A legal target

The Oregon Housing Needs Analysis sets a binding production number of 29,522 homes a year, and every city now has to plan against it. That is the demand aggregation problem already half-solved, in statute.

Patient capital

A public pension fund of national scale, with the duration this kind of asset needs and a plausible in-state mandate.

A freight corridor

An existing rail line down the middle of the state, and companies now building autonomous battery-electric rail vehicles for exactly this kind of regional freight. See the rail corridor.

Robotics depth

A robotics research lineage at Oregon State, a semiconductor cluster to the north, humanoid manufacturing already in Salem, and two federal Tech Hubs. The machines that would automate this factory are the same stack. See the innovation district.

Mass timber

A federal Tech Hub, the TallWood Design Institute, and a materials industry already in the state. The thing being manufactured can be grown here too.

The honest caveat: Oregon manufacturing employment was about 178,900 in 2025 and fell by roughly 10,000 in the year to June 2026. That is a shrinking industrial base. It is a reason to act rather than a reason not to, and it is the strongest argument in any federal application.

Along the housing value chain	Jobs in Oregon	Average wage	Note
Construction	118,200	not published	Near a record high statewide, though down 3.8% in the Salem metro over the past year.
Transportation and warehousing	71,600	not published	Serves all freight, of which building materials are one part.
Wood products manufacturing	22,400	$68,275 to $79,413	Veneer and engineered wood is 8,553 of it. This is the input to panelized building, already made here.
Forestry and logging	8,787	$65,530	Mostly rural, with $717M of payroll.
Support activities for forestry	6,024	$75,884	Growing, up 37% over five years.
Adjacent workforce	roughly 227,000		Our own sum, and an honest one only as an order of magnitude. Construction and freight serve far more than housing, and wood products serve furniture and export too.

The Strategy

Could Oregon become the place where mid-volume works?

The goal we would propose is neither a factory nor winning one company, but removing the constraints that killed the category, so that companies in this band have a structural reason to be here. Housing is the anchor demand because it is the largest and the most urgent, but the same conditions serve robotics, farm equipment, building systems and mobility.

1. Aggregate the demand

Convene the buyers who already have obligations: housing agencies, universities, health systems, cities, and developers with pipeline. Convert scattered intentions into a multi-year order book a manufacturer can finance against. This is the move that decides whether anything else matters.

2. Make components, not boxes

Panels, cassettes and kits ship densely on standard trailers and reach the whole corridor. Whole modules do not leave the Pacific Northwest. The product decision is also the market-size decision.

3. Fix the draw schedule

Stand up an instrument that lets lenders advance against verified factory progress. Until this exists, every manufacturer in the state finances its own work in progress and most of them run out of money doing it.

4. Market the code advantage

Oregon's statewide prefabrication approval and its reciprocity agreements are a genuine competitive asset that could be marketed much harder. Extend the reciprocity list and put it in front of every manufacturer considering a West Coast site.

5. Match the capital to the asset

Pension and insurance capital, structured for thirty years rather than seven. The failures in this sector were funded by money in a hurry, and the successes were not.

And then automate

With utilization solved and capital patient, robotic automation becomes worth its cost. The jobs move from framing crews to controls, maintenance, logistics and the building of the machines themselves, which is the same work the rest of the corridor already does.

The Test

What would have to be true

The conditions that would sink this are published below. Each of these is checkable, and any one of them failing changes the answer.

Aggregated demand has to be real rather than rhetorical. A letter of interest is not an order book. The test is whether public and institutional buyers will commit to multi-year volume, and if they will not, this is a conventional regional manufacturer serving Oregon and Washington.
Component shipping economics have to hold across the corridor. Panels travel further than modules, but not infinitely. If the practical reach turns out to be Oregon and Washington only, the addressable gap is about 33,500 homes a year rather than 245,000. That is still a real business and a much smaller idea.
Autonomous freight has to arrive within about five years, not fifteen. It is the one variable that has genuinely changed since the last generation failed, and it is the one that puts California inside the circle. Everything we claim about corridor scale depends on it.
The state has to use two instruments it already holds. The prefabrication code advantage and a pension allocation both exist as powers today. Neither requires new legislation to explore. If neither gets used, the rest is a private company solving private problems.
Cost has to actually fall. McKinsey put the ceiling at more than 20 percent savings at scale and called it the exception rather than the norm. Careful independent analysis finds little to no saving at typical volumes. We should expect no saving until utilization is proven, and treat any earlier claim as marketing.

Where This Sits

The district as the demonstration

The district is the proving ground

Edgewater in West Salem, and potentially the east bank across the river, is where the first homes and the first shared facilities would go, and where the systems can be tested by people actually living with them. See the District Vision.

The corridor is the market

Interstate 5 and the rail line beside it run from British Columbia to the Mexican border. Salem sits within an hour of everything on the northern half and within reach of the rest if freight economics change. Oregon is where this would be based and where the enabling conditions sit. It is a mark of origin rather than a limit on who it serves.

The state is the enabling layer

Four of the five constraints can only be addressed by a state, which is why these notes are addressed to Oregon rather than to investors. The instruments are the code program, a pension allocation, aggregated public demand and the freight corridor.

The Near-Term Ask

Fund the study first

Nothing on this page is a proposal to build anything yet. What we are looking for now is grant and other early funding to test it: a feasibility study covering both banks of the river in Salem, the innovation district concepts, and whether a mid-volume manufacturing base in Oregon holds up against real numbers on demand, transport, code, finance and cost.

That study is worth funding whichever way it comes out. If the argument holds, it is the basis for a great deal of housing and for a housing manufacturing industry in Oregon. If it does not, Oregon learns cheaply why, and so does everyone else working on the same problem.

Working notes, shared for discussion. Figures are sourced where stated and our own arithmetic where stated. Nothing here has been committed to by any agency, institution or company named, and the scale of anything built would depend entirely on what the study finds.

The Feasibility Work

What it would take to test any of this

Everything on this page should be studied by people with no stake in the answer before it is believed. Three pieces of work would test it, and they are worth funding whichever way they come out. If the argument holds, it is the basis for a housing manufacturing industry in Oregon. If it fails, Oregon learns cheaply why, and so does everyone else working on the same problem.

1. Market and workforce

Who lives where now, what share of the advanced-industry workforce commutes in, what would move them, and what housing types and rents the corridor supports across income bands. The one piece we have actually priced.

2. Site and district feasibility

Both banks of the river: title, environmental, geotechnical, utilities, stormwater, transportation and entitlement pathway, then a concept framework and a financeable first phase. Findings sorted plainly into no material issue, solvable at a known cost, or potential fatal flaw.

3. Mid-volume manufacturing

The question this page raises, tested: whether aggregated demand can be assembled, what components actually cost to move down the corridor, whether the state code pathway and a factory-progress lending instrument work as described, and what any of it does to cost per home. See also the innovation district. Unpriced, because we have found nobody who has done it.

What it might cost, and how much of that we actually know

Being straight about this matters more than the total. One line below is a real quote. Everything else is our own build-up and has not been priced by anyone who would do the work.

Workstream	Range	What it settles
Independent market study	$20K to $30K	Demand, rents, absorption, product mix
Civil and infrastructure feasibility	$35K to $60K	Utilities, access, stormwater
Environmental, Phase I	$10K to $20K	Contamination risk
Geotechnical and site conditions	$15K to $30K	Constructability and foundation risk
Survey, title and easements	$15K to $25K	Assemblage and property constraints
Concept framework plan	$40K to $75K	Capacity, blocks, public realm, phasing
Public and private infrastructure strategy	$20K to $35K	Urban renewal, system development charges, responsibilities
Economic and fiscal impact	$20K to $30K	Jobs, tax base, housing outcomes
Development feasibility and underwriting	$20K to $30K	Costs, residual value, phasing
Legal and entitlement strategy	$15K to $25K	Approval pathway and risk
Program management and grant administration	$20K to $40K	Coordination and reporting
Indicative total	$230K to $400K	Working target of roughly $300K to $350K

The one grounded figure. An independent market consultant quoted a fixed fee of $19,200 for a defined residential market study of the Salem area in August 2026, which is where the $20K to $30K line comes from. That scope covers one site rather than the corridor, so widening it to the questions on this page would move it upward. Every other line is a planning estimate drawn from comparable scopes, and should be treated as a budget placeholder until quotes exist. The third workstream above is genuinely unpriced, because we have found nobody who has done it.

How the jobs number could be built

The first question anyone asks is how many jobs, and we would rather show the method than assert a figure. A number produced before the study is a number produced by whoever wants it to be large. So the economic and fiscal impact workstream above has a specification rather than a target.

Direct employment, from space and density

Square feet delivered by phase, multiplied by employment density for the actual use. Roughly 250 square feet per worker for office, 400 to 500 for laboratory, and 500 to 800 for maker and light industrial space. This is arithmetic rather than forecasting, and it is only as good as the absorption assumption underneath it.

Total employment, from published multipliers

Direct jobs run through regional input-output multipliers for the specific Oregon sector codes, using the Bureau of Economic Analysis RIMS II tables or IMPLAN. For traded-sector manufacturing these typically land somewhere between two and two and a half total jobs per direct job. Higher figures circulate in this field and should be treated with suspicion.

A recruitment ladder, not a single case

Three scenarios with named trigger conditions: slow, where space fills at regional averages; expected, where an anchor commits and pulls suppliers; and strong, where several employers arrive together. Each carries its own absorption curve, and the spread between them is more informative than any midpoint.

Two honesty tests we would insist on

Construction employment counted separately and in person-years, because it is temporary and mixing it into a headline inflates everything. And for the manufacturing question, net rather than gross: factory work substitutes for some site labor, so the study has to say what is genuinely additional and what has moved.

What the factory side actually looks like, and it is smaller than people expect. The most automated volumetric plant in the United States runs about 200 workers against a design capacity of 1,600 modules a year. A California plant stated capacity of two to three thousand units at 150 to 200 employees. Those ratios put a serious factory somewhere near a hundred direct jobs, not a thousand. The larger employment sits around it: suppliers, logistics, installation crews, and the people who build and maintain the automation. Any credible study has to count those separately and say plainly which are new to Oregon and which have moved from somewhere else in it.

And a district comparable for the other half. Cortex in St. Louis reports 5,400 people employed and 406 companies on a 200-acre former brownfield, roughly two decades after five anchor institutions pooled about $29 million. Real, large, and slow.

What would actually be made

Not one product. A kit of parts that assembles into several building types, so a single production line serves a townhouse street, a courtyard block and a mixed-use corner rather than needing three factories. These are design studies rather than finished plans.

A street of timber rowhouses — RowhousesOwnership and family-sized rental, on the streets that connect the blocks.

Aerial view of a courtyard housing block — Courtyard blocksFour to six stories around a shared garden. The workhorse type, and the one that carries the density.

A mixed-use corner building — Mixed-use cornersGround floor for a shop, a clinic or a workshop, homes above.

A slender timber residential tower at dusk — Tall timberWhere land is scarce enough to justify it. Included to be argued with.

A daylit single stair in a residential building — Single stairThe code pathway that makes small floor plates work and gives most homes two aspects.

A street of mixed-use timber buildings with an archway through to a courtyard — The streetHow the types assemble: ground floors that do something, homes above, and a way through to what is behind.

Components of a building kit of parts — The kitPanels, cassettes and cores that ship flat on standard trailers.

A timber component production line — The lineWhere the awkward volume band gets solved, or does not.

Who could hold the money

Now City cannot be the applicant for most of this, and that shapes everything. Federal and state programs will not pay a private developer to study its own site. The eligible applicant is a city, a county, an economic development district, a university, a tribe, or a nonprofit acting with a local government. So the first task is sponsor recruitment, not grant writing.

The good news is that the host does not have to be a charity. EDA's rules make eligible any public or private nonprofit acting in cooperation with a political subdivision, with no 501(c)(3) requirement, which puts regional economic development organizations squarely in scope. SEDCOR, the Mid-Willamette Valley's economic development organization, is the obvious candidate to talk to first, and it has done both halves of this before: it held an EDA Build to Scale award for the Northwest Ag Innovation Hub, and it convened the Newberg workforce housing consortium that secured a $3M state allocation in 2022 and has administered it as a fund ever since. That is a model to learn from rather than reinvent, and whether SEDCOR wants the role is their decision to make, not ours to assume.

Two eligibility limits worth naming early. EPA Brownfields assessment grants were opened to nonprofits by the 2018 BUILD Act, but only to 501(c)(3)s, so that application needs a charity or a public body as applicant rather than a trade association. And Oregon's transportation and growth management grants exclude nonprofits entirely; only cities, counties, tribes, councils of government and special districts may apply.

And a practical one that shapes the ask. Federal awards reimburse after money is spent, so any host carries the cost until it is repaid. For a lean organization that is real exposure, which is why an advance structure or a state appropriation is usually the workable route, and why the ask should be sized to what a host can comfortably carry rather than to what we would like to spend.

Two rules we would hold to whoever hosts it. Describe consultant roles rather than naming firms before an award, because federal procurement standards govern how a public grantee selects its team. And check pre-award cost rules before engaging anyone, because most programs will not reimburse work that started before the award landed.

The doors that are actually open

Verified against program sources in August and September 2026. Dates move, so each is worth confirming before it is relied on.

West Salem URANow · City applies

The city's own urban renewal opportunity fund runs to roughly $50,000 to $100,000 a year and sits on the exact corridor. Smallest paperwork, fastest answer, and the relationship matters more than the money.

EDA planning and technical assistanceSpring 2027 · A regional partner applies

Feasibility studies are the textbook use of this program, and the applicant does not need to be a charity. The prerequisite is getting the project onto the region's comprehensive economic development strategy list, which is slow and should start regardless of anything else.

EPA BrownfieldsAutumn 2026 · City or a 501(c)(3)

Community-wide assessment grants fund reuse planning as well as site assessment, and the riverfront industrial legacy is the hook. Note the charity requirement, which narrows who can hold this one.

Oregon TGM planning grants2027 cycle · City only

Integrated land use and transportation planning, delivered as consultant services rather than cash. Cities choose their candidates months ahead, so the work is getting into Salem's internal queue this autumn.

Business Oregon and Regional SolutionsRolling · A door, not a program

The state's regional coordination function convenes agencies around one project and finds the money that fits. Eligibility varies by program, so this is a conversation before it is an application.

The 2027 legislative sessionGroundwork now

A member-carried funding request is a real lane in Oregon and it is the one that has already worked here: the $3M that funded the Newberg consortium came this way in 2022. It runs on relationships and session timing rather than a deadline, so it starts a year ahead.

PhilanthropyNeeds a charity applicant

Oregon's statewide and family foundations fund predevelopment and planning readily, but the recipient has to be a 501(c)(3). Until an established one is involved, a fiscal sponsor is the interim route.

And an honest limit. Grant timelines can outrun a project. If nothing is executable within a reasonable window we would fund the market study ourselves and proceed, because a study that arrives two years late answers a question that has already moved on.

Sources: EDA eligibility at 13 CFR 300.3; EPA Brownfields eligibility under CERCLA section 104(k) as amended by the 2018 BUILD Act; Oregon DLCD and ODOT for TGM applicant rules; USASpending.gov for federal award history; the market study fee is from a proposal dated 11 August 2026.

What the three states say they need, against what they build

Making things in the thousands

Below this band

Mid-volume, the band in question

Above this band