The Rise of Shippable Microfactories | Hacker News Reader

The Rise of Shippable Microfactories

thesisdriven.com

93 pointsmhb27 comments

Contractors that produce standing metal seam roof panels on site are right inline with this concept. They bring a magic machine with a long roll of steel. Then they form and cut the panels to lengths based on the job site. It’s a super cool process to watch and very efficient.

Same with gutters, for a long, long time. My parents built their house nearly forty years ago, and the day the gutter guys showed up with their machine was young-me's favorite day of the whole project.

There's a video of the robotic building woodworking system for building components.[1] (This article is from their VC firm.) The video is so bad that, if weren't from ABB, a legit robotics company, it would clearly be a scam. They never show the robot at work for more than 1-2 seconds at a time. It's not clear how much of the job it does. The video is about 90% filler - quick cuts, talking heads, scenery. It's not even clear what they're making. All they show is a robot moving some wood around.

Assembling building panels with robots in factories with has been done, but it takes a bigger factory, because what's being made is large.[2] That's a legit video showing the whole process of making housing panels.

[1] https://www.youtube.com/watch?v=qMC-Ls0vI14

[2] https://www.youtube.com/watch?v=O3al52UWoc0

I visited Re:frame Systems[1] in Boston last month and it was pretty cool to see their system which was apparently primarily comprised of a series of magnetic stops on an angled work-surface tended by a giant robotic arm. Cut to length lumber would be fed in on batched trolleys and the arm would compensate for non-square lumber while pinning individual pieces in place then finally nailgun them together in to a completed structural frame. These frames then moved flatpack-style to a second site in the factory for assembly then manual fit-out. I was quite surprised at how slow the arm was, but it didn't really matter because a human would likely take longer by the time they went up a ladder a few times, had a break, and took time off for weather interruptions. They had a pretty impressive team drawing experience from most of the larger US robotics applications and were moving from single level to double level housing, with more elaborate structures planned in future.

[1] https://www.reframe.systems/

Another of the companies mentioned, Facit, has a lot more videos and information about their build process:

https://www.facit-homes.com/building-process

Facit CEO is a friend of mine, really glad they are getting featured here and people appreciate their approach.

The "on site printing" robot looks very practical. A simple little wheeled robot with good navigation drives around on the concrete slab and draws lines and marks to show where things go.

The economics aren't a mystery, people build concrete plants on site for large buildings for the same reason. That said, it makes me wonder if there are things you could do with this sort of approach that would be even more daring. You would be unlikely to fabricate glass on site, but glazing should be doable. Making the windows on site could save time and transport costs. Duct work? We had gutters done which were rolled and welded on site. I could imagine ducting could be similarly created. Custom cabinetry would be another thing I'd wonder about. We've got a neighbor that does custom cabinetry in his 2 car garage, but that setup could be reproduced in a couple of containers. Not that you're going to put that on site at a remodel and block the street, but if you would doing a full housing development? And if you didn't need to lease space for your factory when it wasn't producing would that make for better economics?

In some places in the world (like Pakistan), it is common for upper middle class people to have custom cabinetry in their homes. Usually, during house construction, the carpentry team shows up with the necessary cutting machines. Usually a back of pickup van is sufficient.

Everything is cut from full sized boards as needed on site and installed right there.

It works because wood working machinery is fairly cheap, not heavy and works off standard electric power.

And labour is cheap.

In other parts of the world, you standardize cabinetry sizes and just build for that.

Then a long list of providers offer cabins that fit.

Custom sounds expensive and hard to replace when needed -- what's the upside?

The upside is that people can express their artistic visions by having ornate woodwork in their house. Obviously only available to those rich enough.

But for them, better than https://www.youtube.com/watch?v=t3_ug-IGBJY

When cabinetry (custom or otherwise) is manufactured, it's done with large CNC machines and then assembled.

Unless you were doing a _lot_ of cabinets, the labor cost of an idle cabinet maker overwhelms the shipping cost.

On-site, you're paying the cabinet maker to sand the wood.

https://www.youtube.com/watch?v=WGEF1T3omec

Glulam beams might be manufacturable on site—and it might even be better, since you can make very long glulam beams, but long beams are more expensive to ship.

Most factories have lots of utility connections to their equipment -- kilovolt mains power, compressed refrigerants, toxic waste outputs which must be chemically neutralized before disposal.

Connecting and disconnecting these utilities requires special skills and doing it wrong has major consequences. Special licenses and permits are almost always required by government regulators.

It's the "HVAC guy problem", writ large.

That's true on grid connections, but containers full of batteries could probably overcome that problem.

Porsche has a 2.1 MWhr battery trailer that can deliver 3.2 MW peak they bring to electric track days to keep all the cars charged. That's enough power for some serious industrial processes, and if more is required horizontal scale is pretty simple.

A quick trip overnight to a high powered car charger could fill the batteries by morning. For many loads a simple grid connection constantly trickle charging would be enough power to operate indefinitely.

This sounds... surprising. I thought we were collectively very good at logistics, and that transport was comparatively cheaper than most things. That's why chicken meat takes a round-trip from the US to China and back to the US, because the labor arbitrage saves more than the transport costs, right? That's back and forth between hemispheres!

The examples mentioned -- house framing, gutters -- are a specific category of large weirdly shaped objects. Chicken meat is, in comparison, a liquid. That might explain things.

Factorio Automation 1 research finished.

IRL inserters still cost too much

Reminds me of Nadia traveling all over Mars after the first failed revolution and building habitats from scratch using the huge robots in KSR’s Mars trilogy.

Very cool, but probably not solving the factors involved in housing unaffordability.

In’t the size of the container a limitation?

In the UK, prefab house panels can often larger than a container and constructed on large platforms

The container is used to ship the tooling, you could purchase the raw material local to site and build what you need there.

Sorry, obviously wasn’t clear enough… the jigs to build the panels on are larger than the containers

https://www.facit-homes.com/building-process shows a house being built of boxes made from CNC-cut plywood. The trade-offs are probably a little different than the "whole wall at a time" prefab panels; a little more local labor is required to assemble, but transportation burden is a single delivery/pickup of the microfactory and a supply of plywood instead of shipping the wall-size panels that need a large crane to move.

A nice concept - but at the end, it's still a structure made out of wood and cardboard, and if the staff isn't trained, well just look up Cy Porter and his videos of home inspections for the utter idiocy and cost-cutting [1]...

[1] https://www.youtube.com/@cyfyhomeinspections

With a few hundred highly trained consultants spread through major markets and working directly with contractors with decent jobflow to train them, it could work.

While I'm sure lower construction costs would provide some positive benefits for society, all I see when I look at these sorts of efforts to transition to factory built modular housing is yet another massive wealth transfer from the working class to the capitalist class.

I live in a relatively rural place in Canada where residential construction is on of the primary drivers of the local economy and one of the last bastions of well paying local jobs. You start shipping in a bunch of these "Microfactories" or transition to prefab construction done offsite in some "Megafactory" somewhere, and you put a heck of a lot of people who are supporting families and spending their incomes in my local community out of work.

Rather than a whole town's worth of construction workers swinging hammers making a decent local buck, you now have a scant few foreign billionaires capturing most of the profit in the residential construction value chain. We're being sold the concept on the bullshit promise "Cheaper houses" but all we're going to get is the enshitification of the construction industry, as like with everything else, once startups and VCs get involved in an industry, a few people make billions, but everyone else ends up worse off.

The framers, dry-wallers, concrete pourers, etc. aren't "Retraining" to be "AI prompt engineers" or whatever bullshit we're being fed will be the "Jobs of the future" either, they are going on government assistance or taking minimum wage retail jobs, never financially recovering from the loss of stable, relatively well paying local jobs, and ultimately pulling their families and entire communities down with them.

Meanwhile 1/100th as many people as have been put out of construction jobs around the world are now employed in a few factories owned by billionaires, who then use their wealth to influence public policy to engage in regulatory capture of their industry, and entire swathes of communities become soulless cookie-cutter corporate suburbs of identical, shoddily built pre-fabricated "Homes" that were designed to maximize corporate profits, not for people to actually enjoy living in.

I see the writing on the wall here, as between robotics / automation, economies of scale and VCs pouring billions into businesses that will operate at a loss while they gobble up market share, I definitely think we're going to see a pretty rapid transition to factory built pre-fabricated homes in the next couple decades, as traditional local scale builders just aren't going to be able to be cost-competitive, but from where I'm sitting this is a terrifying rather than hopeful proposition, as at least where I live, it's not going to make housing any cheaper, but its definitely going to put A LOT of people out of work, and we don't really have anything else for those folks to do, so while it is almost certainly inevitable, the transition to pre-fabricated construction is going to absolutely decimate my community and a great many communities just like it around the world.

Reply on news.ycombinator.com