AboutInsightsToolsSign in
All insightsHousing construction
Oct 12, 202618 min

Can Factory Built Apartments Make Housing More Affordable?

A room can be finished before the ground beneath it is ready.

The promise of apartment factories is real. Making that promise reach a tenant requires much more than a production line.

Begin the story
Architectural cutaway showing a finished apartment module aligned above a partly assembled building.

An apartment waiting for its address

Imagine an apartment taking shape inside a factory. The kitchen has cabinets. The bathroom has plumbing. A window already frames the view, although the view is still the other side of the production hall.

Elsewhere, the ground that will carry this apartment is being prepared. Ideally, the two jobs finish together. A truck delivers the module, a crane lifts it into place, and work that would have happened in sequence has happened in parallel.

Now change one detail. The site is delayed.

The apartment cannot move in. It needs somewhere to wait. Someone has already bought its materials and paid the people who assembled it. Interest does not stop because the room has no address yet.

This imagined problem gets us closer to the truth about factory-built housing than the familiar video of apartments being stacked into a building. The crane is impressive. The coordination that makes the crane worth hiring is the harder achievement.

On October 7, 2026, the Department of Housing and Urban Development announced a plan to widen the possibilities for manufactured housing. Among its ambitions is a federal construction framework for larger factory-built residential buildings. That could become an important change in the way apartments are delivered. It is not the invention of modular apartments, which already exist. HUD announcement

The useful question is therefore not whether a factory can produce an apartment. It can. The question is whether enough of the surrounding system can be organized to make that apartment cheaper to deliver, safe to occupy and affordable to live in.

A building mechanism / not a measured project

A finished room.
An unfinished address.

Finishes and services can be completed before the module leaves the factory.

Schematic assembly. Geometry does not encode cost or scale.

Three different meanings of factory built

First, we need to separate a building method from a legal category.

Panelized construction brings parts of a building to the site as panels: walls, floors or other assemblies. Considerable work still happens at the address.

Volumetric modular construction brings three-dimensional sections. Depending on the design, those sections can contain substantial interior finishes and building services. An apartment may occupy one module or several. Modular describes how the building is produced, not a particular material or architectural appearance.

HUD-code manufactured housing is a distinct federal regulatory category. It has historically combined factory production with a permanent chassis and national construction standards. The chassis is the supporting frame, not a requirement to leave wheels under an occupied home. The current reform concerns changing that framework, not turning every modular building into a manufactured home. HUD construction types

Existing permanent modular and panelized buildings generally follow the building codes applicable where they are installed, with off-site approval and inspection arrangements. They do not become exempt from building standards because the walls were assembled somewhere else. International Code Council and Advanced Building Construction Collaborative

The distinction matters to the economics. A producer working under different approval arrangements in different places has a different business problem from one able to repeat a certified system across a broad market. Both may have a factory. They do not necessarily have the same opportunity to use it.

One phrase / three different things

A method is not
a legal category.

Parts of the building

Panelized

Walls, floors and assemblies arrive as panels. More of the building is completed on site.

Three-dimensional sections

Volumetric modular

Rooms or larger sections can arrive with finishes and services already installed.

A federal regulatory category

HUD-code manufactured

National construction and safety standards. Not a label for every modular building.

Schematic distinctions. Modular and panelized buildings generally follow applicable state/local codes. A factory does not remove site or building requirements. HUD construction types.

What HUD is actually proposing

The October action plan follows the ROAD to Housing Act's removal of the permanent-chassis requirement. It sets out work on chassis-free construction, installation and foundations, multi-module designs, energy standards and entry into the federal system for additional manufacturers.

Its larger-building ambition is particularly important: developing a performance-based construction and safety code for residential structures containing five or more dwelling units. The plan identifies common corridors, stairs, accessibility, fire separation, shared systems and escape routes as issues that existing manufactured-housing standards do not fully address. It explicitly describes a framework for further work rather than predetermined regulatory outcomes. HUD action plan

There is an earlier step in this story. HUD announced a rule in September 2024 permitting manufactured homes with up to four dwelling units, alongside fire-safety requirements. The new five-or-more-unit ambition extends that work to the additional demands of larger buildings. HUD's 2024 announcement

The potential prize is a more repeatable approval pathway. Repeating the physical product is useful; repeating the process by which it is accepted can be useful too.

But a construction framework is not a ready parcel. It does not itself supply a foundation, a sewer connection, financing or a tenant. HUD's homeowner guidance still makes site preparation, local permits, utilities and placement requirements part of installing manufactured housing. The October announcement is no evidence that those questions have disappeared for apartment developments. HUD installation and site guidance

It is possible to make the building easier to approve while leaving the development difficult to complete. That is why the cost story needs a wider view than the factory gate.

A regulatory pathway / not an exemption from development
2024

Up to four dwellings

HUD announced a manufactured-housing rule allowing up to four units.

2026

Five or more: a roadmap

October’s plan calls for a larger-building performance and safety framework. Further work remains.

At the site

The rest of the address

Land, foundations, utilities, installation, permissions and funding still have to meet.

September 2024 announcement and October 7, 2026 action plan. The new plan does not invent modular apartments or establish a completed five-or-more-unit code.

The largest saving may be time

A factory changes the order of work.

In the simplest comparison, conventional construction completes the foundation before erecting the building above it. Modular production can make the rooms while a separate team prepares the site. The elapsed time becomes less dependent on the sum of those two tasks and more dependent on the slower of them, followed by delivery, assembly and completion.

The overlap only helps if the two jobs can be coordinated. A late approval can prevent production from starting. A change in the design can undo work already finished. Delivery access or an unavailable crane can spoil a carefully arranged handoff. Early agreement between the designer, builder and manufacturer matters because changes become harder once production begins. HUD's multifamily construction brief

A developer can benefit from finishing earlier even if the direct price of the structure hardly changes. There may be fewer months of site supervision, shorter equipment rentals and less time with money committed to a building that cannot yet earn income.

For a household waiting for subsidized housing, the calendar has another meaning. Four months saved may be four months with a secure address instead of four months waiting for one. The social benefit need not appear in the contractor's margin.

Yet the stopwatch needs to start in the right place. A claim about time spent erecting modules tells us little about the years spent acquiring land, arranging funds and obtaining permission. Faster construction and faster housing delivery are related achievements. They are not interchangeable measures.

Interactive / two schedules / illustrative months

The saving is in
the overlap.

Conventional reference

18 months
Site
0–6
Build
6–16
Complete
16–18

Factory-built experiment

12 months
Site
0–6
Factory
0–10
Assemble + complete
10–12
0 months
CSV

6 months

earlier than the fixed reference

The site is ready before production finishes.

Invented durations, not typical or measured performance. The conventional 18-month reference stays fixed; the delay changes only the factory site. Factory production finishes at month 10, and assembly/completion takes two months. Land acquisition and permitting are not timed here.

What the evidence says about savings

There is evidence of an advantage, but it is less cinematic than the largest promotional claims.

In a July 2024 account of a Department of Energy-supported study, researcher Kevin Grosskopf described a comparison of 48 modular multifamily projects with 158 site-built projects in several major US housing markets. The modular projects were completed about 30% faster on average, while the reported overall cost advantage was about 5%. National Institute of Building Sciences

Those two figures belong together. A large difference in time can coexist with a much smaller difference in money.

They also belong to that study, not to every apartment building. Project selection, unit sizes and the kinds of housing delivered differed. This was not a randomized experiment that changed only the construction method. It cannot establish that an otherwise identical project in a particular city will save either percentage.

The numbers are useful precisely because they temper the fantasy. A factory may offer a meaningful improvement without cutting the whole development budget in half. Five percent can be consequential on a large project. It just needs to be five percent of the budget that actually exists, not an exciting percentage attached to an undefined slice of it.

Researcher’s July 2024 account / observational project comparison

A bigger change in time.
A smaller change in cost.

30%

Faster completion

Average timing advantage reported for modular projects

5%

Lower overall cost

Approximate reported cost advantage

48 modular projects

158 site-built projects

Different project mixes.
Not a randomized experiment.

Kevin Grosskopf, NIBS, July 2024. These averages cannot establish the saving for an otherwise identical project. Both bars use a 0–35% scale, but time and cost are different outcomes.

Follow the discount through the whole budget

Consider an illustrative development budget of $350,000 per apartment. These are invented inputs for a cost experiment, not a contractor quote or an estimate of a typical US project.

Illustrative budget per apartment
Budget itemUSD
Land allocation$65,000
Design, approvals, financing and other development costs$45,000
Structure and interiors being compared$180,000
Remaining site work and completion$60,000
Total$350,000

Suppose factory production reduces the $180,000 portion by 15%. That is a $27,000 saving. Suppose the modular option also brings $12,000 of additional transport, lifting and coordination costs not included in the baseline.

The net saving is $15,000. The revised total is $335,000, about 4.3% below the original budget.

The factory's 15% reduction has not become a 15% reduction in the whole project. It has become a smaller saving after the unaffected costs and additional expenses are included.

That does not make it a failure. It tells us what was achieved.

Now give the same project an expensive retaining wall, difficult ground conditions or a higher land price. Keep the manufacturing saving unchanged. The total percentage benefit becomes smaller because more of the budget lies outside the factory's reach.

Alternatively, retain a simple site, repeat a proven design and bring the factory closer. The extra expense could fall and more of the production saving could survive. These are different projects, not contradictory verdicts on modular construction.

Whenever someone says a method is cheaper, ask what has been counted. Does the number include the site, financing, delivery and the work needed before residents can occupy the building? Is it a preliminary budget or the final account? Is the comparison between equally sized homes with equivalent performance?

The question is not hostile. It is how a promising method becomes a believable business.

Interactive / follow the discount / USD per apartment

A factory discount.
A whole-project budget.

Conventional budget

$350,000
$65,000
$45,000
$180,000
$60,000

Factory-built budget

$335,000
$65,000
$45,000
$153,000
$60,000
$12,000
Both budgets, item by item
CategoryConventionalFactory built
Land$65,000$65,000
Other development$45,000$45,000
Structure and interiors$180,000$153,000
Site and completion$60,000$60,000
Added modular expense$0$12,000
$15,000

4.3% whole-project saving

$27,000 manufacturing saving, less $12,000 additional modular expense.
15%
$12,000
$65,000
CSV

Illustrative inputs, not a quote. Other development costs remain $45,000, structure/interiors start at $180,000 and site/completion stays $60,000. The manufacturing discount applies only to structure/interiors. Increasing land changes the denominator, not the factory saving.

The money can arrive in the wrong order

Construction lenders have to finance a process, not merely a finished building.

Factory production can require earlier payments for materials and completed modules. HUD's financing brief identifies higher upfront requirements and recommends payment milestones that follow materials, production, transport and installation. It also calls for contracts that allocate storage and transport risks. HUD's multifamily financing brief

The practical tension is easy to understand. The developer wants a reliable production slot. The manufacturer needs money to produce. The lender needs acceptable security and a way to monitor what its money has purchased while the work is still off site.

This is not solved by showing the lender a beautiful completed apartment. It is solved through an acceptable financing structure and clear responsibilities while the apartment is incomplete or waiting to travel.

The timing can change the interest arithmetic. In a separate illustrative example, assume a project finishes four months earlier and would otherwise have carried an average $12 million loan balance during those months at 8% annual simple interest. Avoiding that interval saves $320,000 in interest.

Now assume factory payments mean an additional $6 million is borrowed four months earlier than it would have been under the conventional schedule. At the same interest rate, that earlier borrowing costs $160,000. On those assumptions, the net interest benefit is $160,000, before considering fees, operating income or other costs.

Neither balance describes a measured project. The example isolates why earlier completion is not enough to calculate the saving: we also need to know when the money was drawn.

A faster building can be a better investment. It can also demand more cash sooner. The financing has to accommodate both.

Interactive / two separate interest intervals / simple interest

Finish sooner.
Borrow sooner, too.

Interest avoided at the end

$12,000,000 × 4 months × 8% / year

012 months
$320,000
Rectangle area = balance × time.
Shared scales: $0–$12m and 0–12 months.

Interest added before assembly

$6,000,000 × 4 months × 8% / year

012 months
$160,000
Rectangle area = balance × time.
Shared scales: $0–$12m and 0–12 months.
4 months
4 months
8%
Net interest benefit$160,000

$320,000 avoided − $160,000 added

CSV

Separate illustrative exposure intervals, not a full loan-draw curve. Fixed balances of $12m and $6m; no fees, compounding or operating income. Do not add this study to the per-apartment budget experiment.

A housing shortage does not guarantee a full factory

The country can need millions of homes while a particular production line runs short of orders.

A need is not a purchase order. To place an order, a developer needs a site, a compatible design, permissions and funding that can arrive on a schedule the manufacturer can use.

Imagine a factory with $6 million in annual costs that remain fixed within the range we are examining. At 1,000 completed apartments a year, those costs allocate $6,000 to each apartment. At 500, the allocation is $12,000. At 250, it is $24,000.

This is a simplified overhead model, not an account of an actual factory. It leaves out variable production costs and the possibility of cutting expenses or changing capacity. It shows one important pressure: an idle production line can make the apartments that do get built more expensive.

The producer therefore needs a reasonably dependable sequence of work. The developer, meanwhile, may hesitate to commit before approvals and financing are secure. Each is waiting for a certainty the other cannot quite offer.

A HUD-commissioned comparative study posted in March 2026 identifies regulatory reform and aggregation of demand as prerequisites for accelerating off-site construction. Its international research examines not just buildings but the systems around them, including certification and the organization of orders. Those are the researchers' findings, not a guarantee that copying a foreign program will produce the same result here. HUD-commissioned comparative study

Combining credible projects into a steadier pipeline could help spread fixed costs and learning over more homes. But promising volume that cannot actually be financed creates a different problem: the factory may invest in capacity for orders that never arrive.

Predictability is worth something only when the prediction can be relied on.

Interactive / annual fixed overhead / not total unit cost

The same factory.
Fewer apartments to share it.

Fixed overhead per apartment

$0k$6k$12k$18k$24k2505007501,000

Apartments completed per year

A pipeline, not just a need

50 filled / 50 unfilled slots
One slot = 10 apartments in a 1,000-apartment reference.
500
$6,000,000
CSV

$12,000

fixed overhead per completed apartment

$6,000,000 ÷ 500 apartments
Compare three output levels
Annual outputFixed overhead per apartment
250$24,000
500$12,000
1000$6,000

A simplified fixed-cost model. Variable production costs, resizing the factory and capacity adjustments are excluded. Grid slots are a visual reference, not orders from an actual producer.

America has tried the production revolution before

In 1969, HUD began Operation Breakthrough, a national effort to demonstrate industrialized housing. It selected 22 housing-system producers and built more than 2,900 prototype homes across a group of demonstration sites. The program examined far more than an assembly line, including transportation, performance and site planning. HUD's historical archive

It is tempting to cast that history as a simple failed experiment. The government tried to manufacture housing, the revolution did not arrive, and therefore the idea must have been wrong.

The contemporary verdict was more specific.

In November 1976, the Government Accountability Office concluded that the program had not demonstrated the marketability of most of the construction systems it sponsored. Its explanation included an unexpected deterioration in the housing market, the suspension of subsidized mortgage programs and systems that lacked sufficient cost-saving potential. GAO also credited the program with useful changes in methods, evaluation and building codes. GAO's assessment

Those distinctions matter. A technically workable product can fail to find a financially workable market. A subsidy program can support production and then withdraw the conditions on which production depended. Some technologies can disappoint even while others contribute useful lessons.

History gives us no reason to assume today's attempt will fail. It does give us a reason to demand more than a successful demonstration. Show that the product performs. Then show that someone can finance it, approve it, order it again and receive it at a price that sustains the producer.

A prototype answers whether something can be built. A functioning housing industry has to answer whether it can keep being built.

Operation Breakthrough / a prototype is not a repeatable market

The factory has
a history, too.

1969

A national demonstration

HUD selected 22 housing-system producers. More than 2,900 prototype homes followed.

1976

The marketability test

GAO found most systems’ marketability unproved. Demand, financing and cost potential mattered alongside technology.

HUD’s historical archive and GAO, November 2, 1976. Schematic, not an archival photograph. The assessment also credited useful testing and code changes.

The San Francisco example that deserves a careful reading

Tahanan, at 833 Bryant Street in San Francisco, offers a much more recent example. Its 145 permanent supportive apartments were built using modular units, with Mercy Housing as developer and partnerships involving Tipping Point Community and the San Francisco Housing Accelerator Fund. Mercy Housing says construction finished in November 2021 and the building was fully leased by February 2022. Mercy Housing's completion account

The story is attractive because it involves homes for people who had experienced homelessness, in a city where delivering subsidized housing is particularly expensive.

The Terner Center's February 2021 analysis, written while the project was still under construction, projected a development cost of $382,917 per apartment and a 33-month period from land acquisition to completion. The researchers identified a package of advantages: explicit cost and time goals, flexible capital, streamlined approval and modular construction that overlapped factory work with site preparation. Terner's 2021 case study

A later Terner report, published in August 2024, described the completed development as costing about $377,000 per unit over just more than three years. Terner's 2024 report, page 19

Neither account isolates the factory as the cause of the entire saving. A construction method, an approval process and a financing approach improved together.

There is another detail that deserves attention. Mercy Housing describes upfront philanthropic investment and a subsequent long-term city lease that supports permanent debt service, along with operating and service funding. Avoiding upfront public financing did not mean avoiding public support throughout the life of the development. Mercy Housing

The lesson is not that every city can order apartments for $377,000 in 2026. These were particular units, on a particular site, under a particular arrangement, completed years earlier. The lesson is that production can work much better when permission, funding and design stop pulling in different directions.

The factory was part of that achievement. It did not achieve it alone.

Tahanan / San Francisco / four levers, not one isolated cause

The building was modular.
The achievement was collective.

01

Clear cost + time goals

02

Flexible capital

03

Streamlined approvals

04

Modular production

145supportive apartments

Nov 2021construction completed

≈$377kper unit in Terner’s later account

Mercy Housing, 2022 and Terner, August 2024, p.19. A past supportive-housing project with several advantages and ongoing public support, not a 2026 modular price quote.

Repetition should improve the building without flattening the home

A repeatable bathroom does not require a city of identical buildings.

Some things are worth standardizing because residents benefit when they work reliably: service connections, inspection records and details that keep water out. Other things deserve variation because homes belong to places and people: the relationship to the street, daylight, communal space and the mix of households a building can accommodate.

The design challenge is deciding which differences provide value and which repeatedly consume money without improving anyone's life. That is more demanding than either insisting on a unique solution to every component or accepting the same building everywhere.

Performance also has to be judged in the assembled building. A completed room still meets another room, a corridor and the outdoors. HUD's larger-building proposal recognizes that shared systems and fire separation require attention beyond the individual dwelling. HUD action plan

Nor should a construction label stand in for an environmental assessment. Fewer on-site operations may be useful, but transport, materials and performance over decades still belong in the comparison. The appropriate question is what the complete building uses and how it serves occupants, not whether the construction video looks clean.

For residents, the important tests will be ordinary ones. Can they sleep without unreasonable noise? Does the home remain comfortable? Can faults be repaired by someone who knows what was installed? A faster production system earns its reputation through the years after the crane has left.

Explore the assembled home / schematic, not a blueprint

Repeat the detail.
Keep the life.

Connections

Repeatable service connections and inspection records can make installation and future repairs more reliable.

Cheaper construction does not automatically mean cheaper rent

There is one final handoff, and it is the one that matters most to the household.

Suppose the factory delivers the $15,000 saving from our earlier example. The building costs less to develop. Who receives that benefit?

There are several possible outcomes. A marginal project may become feasible and add homes that would otherwise never have existed. A developer might accept the same rent with a better return. Competition could make a lower rent attractive. A landowner could demand more for a parcel that has become easier to develop.

These are economic possibilities, not a finding about a particular project. The saving is not mechanically attached to the tenant's lease.

An illustrative financial hurdle makes the distinction visible. If a project requires annual net operating income equal to 6.5% of its cost, a $15,000 reduction in cost lowers that requirement by $975 per apartment per year. Spread evenly over twelve fully occupied months, with operating expenses unchanged, that is about $81 a month of additional room in the budget.

It is not a predicted rent reduction. The 6.5% hurdle is an assumed input, not a quoted market requirement; vacancy, taxes, insurance and the financing arrangement could change the result. Market rent does not emerge from this calculation. What it demonstrates is the scale of one possible benefit before deciding how it is used.

For deeply affordable housing, a lower development cost can let a limited capital budget support more apartments. But the money needed to operate the building and bridge the gap between a household's affordable rent and the cost of providing the home remains a separate question. Tahanan's ongoing city support makes that distinction concrete.

We should welcome a reduction in the cost of producing housing without pretending it performs every job that housing policy needs done. A manufacturing improvement can widen the feasible supply. Affordability also depends on access to that supply and the incomes of the people who need it.

Interactive / an illustrative feasibility hurdle / not a rent forecast

Room in the budget.
Not a promise in the lease.

Development saving per apartment
$15,000
Assumed annual NOI hurdle
6.5%
Annual room in the budget
$975
$81.25 / month

Room in this feasibility budget. Not a predicted rent cut.

$15,000
6.5%
CSV

Lower development costs
may reach...

More homes

A previously marginal project can become feasible.

Developer return

The same rent can support a better return.

Lower rents

Competition or a funding condition can pass savings through.

Land value

A more viable development can support a higher site price.

Saving × assumed annual NOI hurdle ÷ 12. Full occupancy and unchanged operating expenses. Possible uses of a saving are not probabilities or measured distributions. This study is independent of the other controls.

An open blue apartment door leading into a sunlit home.
The final handoff

Someone unlocking
a home they can afford.

What a serious apartment factory would have to prove

The next phase deserves an ambitious test, but one with a complete scorecard.

Measure the final development cost, including delivery and the site, rather than only the factory contract. Measure the time from a defined starting point to legal occupancy, rather than the days in which a crane is visible. Compare homes with similar sizes and performance, and distinguish projects that finished from those that stalled.

Follow the production line between projects. Can it maintain a credible pipeline without relying on promises that vanish when funding changes? Follow the financing while the work is off site. Are responsibilities clear if a manufacturer fails, a module is damaged or an approval arrives late?

Then follow the residents. Is the building comfortable and maintainable? Did savings create additional homes, improve long-term affordability or merely move into a different part of the transaction?

HUD's October plan could help make some of those answers better. A more coherent regulatory pathway would give producers more room to repeat and improve their systems. Whether the promise reaches residents will depend on the completed developments and the businesses that keep delivering them.

Return to the apartment waiting in the production hall. Its kitchen can be finished before the ground is ready. That is an extraordinary capability, provided the ground will be ready when the apartment needs it.

The task is to bring the room and the address together, repeatedly, without losing the saving on the journey. If apartment factories can do that, the most persuasive image will not be a module hanging from a crane.

It will be someone unlocking a home they can afford.

Sources and study notes

Research checked October 12, 2026. The October 7 HUD action plan is a roadmap for further regulatory work, not a completed five-or-more-unit construction code. Building method, federal regulatory category and local development permission are separate questions.

The reported 30% timing advantage and about 5% cost advantage are attributed to Kevin Grosskopf's July 2024 NIBS account of an observational comparison of 48 modular and 158 site-built projects. Project mixes differed. We did not acquire the underlying project-level dataset and do not infer a causal saving for an otherwise identical development.

Tahanan's February 2021 figures were projections. The later approximate $377,000-per-unit figure comes from Terner's August 2024 report. Construction finished in November 2021; the building was fully leased by February 2022. None is a 2026 price quote.

All five interactive studies are separate illustrative experiments. They are not calibrated forecasts. The timeline assumes an 18-month conventional reference and a 12-month factory schedule with production and site work in parallel. The delay control changes the factory site's readiness; the conventional reference stays fixed. It does not model permitting or land acquisition.

The budget study keeps other development costs at $45,000, structure and interiors at $180,000 and remaining site/completion at $60,000; land, manufacturing discount and additional modular delivery/assembly expense are editable. The finance study uses simple interest on two separate, constant exposure intervals, not a complete loan-draw history; fees and operating income are excluded. The factory study divides assumed annual fixed overhead by completed apartments; variable costs and capacity changes are excluded. The resident study multiplies a development saving by an assumed annual NOI hurdle and divides by twelve. It assumes full occupancy and unchanged operating expenses. It is room in a feasibility budget, not a rent prediction.

The architectural cutaway and doorway are conceptual illustrations created for this article. Chart marks, values, labels, controls and exports are coded from the visible assumptions, not generated inside an image.

Read as Markdown Explore the rental pro forma
More insights