Robots to the Rescue? A Very Human Problem

We have a housing crisis. The median U.S. income earner can afford a home that costs around $298,000, yet the median U.S. home price is $440,000. There is also an overall housing shortage.  Paradoxically, the new homes being built are sitting on the market longer, because they are just too expensive for most buyers to afford.  This isn’t the result of homebuilders making poor decisions.  Many of them would love to service all this pent up demand and churn out affordable homes that would be snatched up by eager buyers.  They just can’t because their costs are too high to build these affordable homes and still make a positive margin.

There are many opinions on the cause of these gaps. We think there is a major underlying root cause. In the 1960s economist William Baumoi proposed his theory of cost disease. It posits that products produced in industries that fail to increase worker productivity through technological advancement (primarily automation) will become increasingly expensive over time relative to products that are produced by more advanced and automated industries.This all-too-well describes the residential construction industry. The homebuilding process hasn’t incorporated much disruptive technology since the circular saw. So, if the problem is that homebuilders haven’t automated, is the solution just to automate? Probably, but it’s easier said than done.

It’s 2025 Already, Where Are the Robots?

Do you think anyone in 1930 would have thought technology would replace their research physicians and chess champions before it replaced their local carpenter? That is exactly what happened. AI has been developing groundbreaking treatments for complex diseases for the past several years with little human direction. And a free app on your smartphone can easily beat the best chess players in the world. Technology in any form has remained conspicuously absent from the construction site. Asking robots to build something in the real world has been a sneaky hard engineering challenge. Robots want to work in the perfect digital. The real world is never perfect. Construction is chaotic. The foundation is never really level, the weather is never the same, the 2×4’s are all warped, and the measurements are always a half inch off from the plans.Tradesmen employ their judgment to adjust for all this on the fly. Robots don’t have judgment. This constant variation blows their rigid little robot minds. We are determined to get these robots to build us some houses. And the industry is finally making some progress.

Build Like a Robot, Not a Human

We think the most promising forms of construction automation let robots do what they do best, and don’t try to make robots build like humans build. This means rethinking the process from the ground up. For example, instead of trying to build a robot that can stack CMU blocks, ask why we are building with CMU blocks in the first place. It turns out a CMU block is about the biggest block a human can easily lift and stack (ditto for a 2×4). Robots can lift much more and move faster and more precisely. They don’t have good eyes, and struggle with balance and spatial perception, so stacking things is hard for them.

Enter 3D printing. Let the robot use its superhumanly steady hand to lay down the wet concrete layer by layer. The fluid nature of the wet concrete can fill in tiny variations. So far, the market agrees. 3D construction printing is gaining traction industry wide, with bigger projects breaking ground every day. Right now, our company is building three separate housing developments (one with 80 affordable homes), a hotel, and several custom homes. And we have a long project waitlist. Other forms of construction automation are lagging. The giant block stacking truck is showing some promise, but it’s expensive and hasn’t completed many projects. The humanoid robots are years away (if there ever will be a good use case for them). There are tradesman robots making the rounds at builders’ shows- like one that installs shingles, one that floats drywall, or one that paints. They have been more notable for YouTube clips of them going haywire on the jobsite than for completed projects so far. We are starting to see some automated heavy equipment come to market to do sitework, but these machines are still rare and expensive.

Nothing Worthwhile is Easy

As 3D construction printing gains traction, its drawbacks are also coming into the light. The main problem has been cost. The whole point of this was to address housing affordability. So, it seems pointless to build robots that build houses that are more expensive. Others in the space have focused on selling high-end homes for premium prices, because 3D printed homes are often more durable and energy efficient, and can be architecturally stunning, and can thus command premium prices in some markets. Many 3D printing companies may be boxed into these higher prices for myriad reasons. Their 3D printing mortars may cost too much, their construction printers may be too expensive, print too slow, or require too much expensive infrastructure to mobilize, and they may require too many highly compensated engineers to operate.  There are systems emerging that have smaller footprint, and that print faster with lower cost materials that are breaking through these cost barriers.

We think the key is the material science, which is tougher to get right than the robotics. The ideal 3D printing mortar is carbon friendly, super-structural, and sets fast so the robot can print fast. Most importantly, it’s low cost.  In short, if the materials that a 3D printer uses to build cost more than the materials a traditional construction worker builds with, saving money on 3D printing is going to be difficult.  Working with such fast-setting materials is challenging and requires custom-built mixing equipment, but we think it’s the only path forward. 

The Right Robot For the Right Job

We love 3D construction printing, but we also know it’s not a complete solution. We love concrete, even though it doesn’t make sense to make everything out of it (a great wall, not so great mattress). We think prefabrication compliments 3D printing well. Robots can work offsite in a factory making some of the house, while robots work onsite to make the heavy parts that are expensive to transport. We print 3D print walls, showers, counters, landscaping elements, and even outdoor furniture. And we have developed a structural insulated panel system that we use to make roofs, non-structural walls, and mechanical, electrical, and plumbing installations. When the robot painters are ready, you will see them on our jobsite too. We think these multi-disciplinary approaches are a key to unlocking cost savings.

The Bridge Between AI and the Real World

Until we all transition to the matrix, AI will need a way to translate all the intellectual work it can do to the real world. If we don’t create robots to do this physical work, we will find ourselves in a world where AI does all the thinking, and we do the low-cost repetitive labor. We don’t think this is a great path for humanity. Instead, we hope the construction industry will embrace automation, so we can give our front-line workers better tools to make their jobs and lives easier and more productive. We welcome everyone to join us on the journey.