Folding at Home
Taking a look at foldable structures
As we’re fond of pointing out around here, prefabricated construction has a long and storied history. Because of this, and because the basic elements of buildings haven’t changed substantially in the last 75 years or so, it’s rare to see truly new and unprecedented building systems. Instead, we see the same basic concepts reappear over and over again.
One of these recurring ideas is the concept of a foldable house. We can find patents for this idea as far back as 1918, and the concept shows up every so often as a revolutionary take on the stodgy and traditional construction industry. We see it on magazine covers in the 1950s, and as conceptual prototypes by the Army in the 1970s.
The basic idea behind the foldable building is simple: the house gets built in a factory as a series of panels connected by joints, sliders, and hinges. It gets shipped to the jobsite folded-up and packed tightly together. Once it arrives, the panels slide out, the walls and roof swing open, and the house unfurls into it’s full size.
Folding Buildings in the Wild
There’s actually quite a few of these systems on the market today, all of them implementing a slightly different folding system, and all of them tackling different niches in the construction market:
Boxabl is probably the most well-known of these, at least in the US They make a folding 375 ft2 ADU that packs into a shipping-container sized package, and can be unloaded and fully assembled in just a few hours. Not yet available for purchase, but you can pre-order one for $50,000.
A-fold is an italian system that can unfold into an A-Frame house
Brette Haus makes a sort of clever folding tiny house, where the entire roof gable folds down over the side into a more compact box.
Ten Fold is a developer of lever-based folding systems. Their website is full of renderings of insectoid-looking houses and other structures that can expand from a shipping-container sized package. Unclear if it’s ever actually been used to build anything, or if it’s all speculative.
Several Chinese companies, such as Dream Maker, or China Containers, and a bunch on Alibaba, make various flavors of folding prefabricated structures. Some of these are shipping container-sized modules with recessed assemblies that slide out. Others have hinged walls that allow the entire structure to flatten. Often used for workforce housing, emergency shelters, or other low-end, temporary accommodations.
EBS-Block, another Chinese company, makes a 600 ft2 home that unfolds from a shipping-container sized module
Flexotel makes foldable accommodations designed for concerts, festivals, and other outdoor multi day events.
Different branches of the military have various expandable shelter solutions.
In addition to these, we often see folding elements implemented as part of a more traditional prefab system. Broad Group uses folding floors as part of their new building system, and manufactured homes often have a hinged roof that can be lowered during transportation.
Foldable building systems vary quite a bit in the amount of services they come from the factory installed with, and thus the amount of site work they require once they’ve been unfolded. Some, like Boxabl, are basically completely turn-key and can be placed and set up in a single day. But some take far longer. A-Fold’s system, for instance, seems to include just the structure, and it takes another 45 days of site work to add finishes, etc. Blu Homes folding home reportedly took 3 months to complete on-site.
Theoretically, this method of construction takes the best features of both panelized and volumetric modular construction. Since the panels are already connected together, the building requires very little work to assemble once it’s arrived on site. This solves the issue of how you field-connect high level-of-completion panels, and reduces potential alignment problems (which are nasty to field-fix). And because it folds up, it can be transported much more compactly, and thus much less expensively, than volumetric modular. Foldable buildings thus allow the level of completion of volumetric modular, with the shipping efficiency of panelized construction.
The bear case for foldable buildings is that the folding mechanisms add so much expense to the building that they negate any potential savings. A building by design is a static, unmoving structure made from the least expensive materials possible. Adding folding capabilities means introducing substantial new loads and load paths on portions of the building that wouldn’t normally be loaded (a wall isn’t normally designed to have another wall hanging off of it, supported by hinges). This inevitably requires heavy structural hardware, and often complex assemblies to ensure the folding elements can be moved safely and easily. As a result, folding systems mostly seem to be made of steel, and often have cables, wheels, heavy hinges, hydraulics, and other parts that you wouldn’t normally find in a building (such as in the EBS-Block building above). These sorts of mechanisms are most visible on systems like Ten Folds, but even Boxabl’s construction is much heavier-duty than what you’d normally get in an ADU. The bigger the building you’re trying to build with a folding system, the worse this problem gets.
For the bull case for folding buildings, we need to detour a bit.
Buildings and Pace Layers
We’ve talked before about Stewart Brand’s “Pace Layer” concept, where a building can be broken down into several separate layers:
The original idea of pace layers relates to system aging. Each one of the above systems will age and be replaced at different rates: site and structure last the entire life of the building, but the space plan (the layout of the rooms) might change every 10 years or so. Because of this, care should be taken to make these systems as independent from each other as possible.
These layers make up the different functions of the building. Functions aren’t contained in one particular facet of the building, but extend throughout it. A building is a machine for controlling an interior environment, and that environment needs to extend everywhere. The waterproofing can’t simply be contained in some “waterproofing” subassembly tucked away in a corner, but needs to encompass the extents of the building. Same with the plumbing, the mechanical services, the finishes, etc. This is one reason why even modular construction isn’t very modular
This layering concept also applies to how a building is built - a building is built a layer at a time, in a very similar fashion to Brand’s layer breakdown:
Before construction can begin, we need to perform the site work - grading, compacting, building the foundation, running the service lines to the building, etc.
After that comes structure. This consists of the walls, floors, roof, sheathing, and other elements that physically support the building.
This is followed by the skin - the windows, water barrier, and similar systems to make the structure watertight.
Once it’s been made watertight, we can proceed with services - electrical, mechanical, and plumbing systems, gas lines, etc.
Simultaneous with services, we’ll have the installation of the outside finishes - architectural cladding, roofing, etc
After services, we have interior finishes - drywall and flooring, followed by plumbing fixtures, cabinetry, lights, etc.
Last will be the installation of any stuff that’s part of the construction process - appliances, etc.
This is one reason why buildings are so time-consuming to produce: there’s often limited ability to do this work in parallel. Most of these layers can’t be started until the previous layer has been finished - I can’t build my structure until my foundation is complete, can’t put in services until the skin is complete and the building is watertight, can’t put on drywall until the services have been installed, etc. This becomes less true as your building gets taller (where structural work might be in process on the top floor while the services and finishes are being installed on the lower floors), but most buildings aren’t tall enough to leverage this.
Because both its construction and its functions exist as a series of layers, in some ways the most natural way of conceptualizing a building is not a 3D object at all, but a 2D object that has been folded into the third dimension .
This style of assembly isn’t uncommon. The most central example is of course origami, where a 2D sheet of paper is folded up into a 3D object (such as, say, a house). But we see it elsewhere. Clothing is made from 2D sheets of fabric folded and connected to make 3D garments. Proteins start life as 1D chains of amino acids, and get their function from bending and coiling into complex 3D shapes. In the construction world, light gauge steel sections start life as 2D sheets of steel that are then bent into useful shapes.
This perhaps explains why we have such a hard time assembling buildings efficiently. Assembly strategy should align with an object's geometric orientation, and our methods of building mostly don’t take advantage of this layer-based arrangement. It's as if we tried to build a protein by putting together 3D chunks to try to get the right shape, or tried to create a shirt by sewing 3D tubes for the arms.
Considered this way, a folding structure seems like the most natural way of assembling a building. Taken to the limit, it suggests buildings should be built by flattening them out as a single 2D surface, adding the layers one by one, then folding the whole assembly into the shape of the building at the end.
A building constructed like this actually resolves a number of thorny construction issues:
Automated building assembly becomes much more feasible. We’ve speculated as to whether 3D printing will ever be able to print more than a small fraction of a complete building, but it’s easy to imagine how even a fairly simple gantry system with some kind of tool-changer could build up a flattened 2D building layer by layer (another free startup idea for someone).
It enables a greater degree of simultaneous construction on-site. Instead of needing to wait for the structure of one floor to be done before starting the next, all your levels could be built simultaneously.
Inspection becomes incredibly easy (just take a single high-resolution photo of your build site at frequent intervals).
The problem of assembly order, and orienting large pieces (such as sheets of drywall, or bathtubs) to fit inside some enclosed area is completely eliminated - everything can be dropped in from above.
Of course, all that ignores the actual reality, which is that assembling a building as a completely 2D object is ludicrously impractical for any number of reasons:
Spreading out all the surfaces of your building would take an enormous amount of space. A 2,500 square foot house might flatten down into 10,000 square feet of surface area (nearly a quarter of an acre).
If you’re building outside, your building remains exposed to the elements right until it's folded up. And you’d need to create an enormous, flat surface to lay all your elements on.
The problem of heavy hardware required for folding would get dialed up to 11 - every joint (and there would be a LOT of joints) would need a hinge capable of handling an enormous load.
No practical way of folding up a completely flattened building (though I can imagine some severely impractical ways).
I could go on. The idea is sort of a spherical cow, an overly simplified model that abstracts away all the practical concerns.
But simplified models have their place - though we can’t actually build a building like this (yet - growth mindset!), I think it’s still a useful way to think about buildings.
Back to Reality
Where does that leave us with actually existing foldable structures?
In general, both panelized construction and volumetric modular build at a slight cost premium to conventional construction. This seems to hold true for folding structures as well. Boxabl comes in at $133 per square foot before you add in the cost of the foundation, sitework etc. A-Fold comes in at around $114 per square foot. So evidence suggests that these aren’t a more efficient way to build, but are just picking a slightly different spot along the existing cost frontier.
The majority of folding systems are marketed towards portable structures - disaster relief, emergency hospitals, military deployment, workforce housing, concert accommodations, etc. Situations where the stay will be long enough that you need a real building, but that ultimately will be moved elsewhere. They’re designed for repeated transportation. Much of the complexity and expense of their folding mechanisms stems from their need to be repeatedly folded and unfolded.
I suspect a folding system designed for a one-way trip might look different.
You can imagine a folding system designed for this. It could consist of light framed prefab panels that would come from the factory at a medium level of completion - structure, sheathing, MEP and insulation would all be preinstalled, but not drywall or any other finishes. Instead of heavy structural hinges and complex mechanisms, they could be joined by simple light gauge steel angles - something strong enough to survive a single unfolding, basically. The panels would be unloaded, and carefully folded into place by hand (to avoid placing undue load on the structure), after which work would proceed on site in a conventional fashion. It could even be done on-site, similar to tilt-up construction.
This sort of system gets you a lot of the benefits of a folding system (reduced sitework, low transportation costs, no alignment problems between panels) without the drawbacks.
 - By 2D, I mean having negligible thickness compared to its dimensions along its other axes, not literally only existing in two dimensions.