Boston Dynamics robots are still nowhere close to as dexterous, flexible, or capable as human laborers. Getting them to build houses is a massive and expensive R&D project, not a startup, for now.

If you built the houses the same way as with human labor, sure. But perhaps they would be built differently with robots, in a way which didn't require that level of dexterity. After all, most industrial robots don't have opposable thumbs.

I'm not saying it's a simple problem to solve, nor one whose solution I have fully mapped out in my head. Just that it's a problem whose solution strikes me as being within the realm of feasibility.

There is this https://www.mudbots.com/

The tiny house they have a gallery of is aggressively ugly, with lots of bizarre curved walls seemingly just to demo their method. And the shelves and fixtures look precarious and unsafe, straight out of Arrested Development. Given the build quality, I'd be afraid the roof would come down on me if I sneezed.

I have Also heard of https://www.14trees.com/ and terran robotics doing this...

Edit: I imagine though, that in locations where building is regulated under the international residential code. There has been less investment, it only in the last few years added an appendix for mud/cob construction... So you can add regulatory cost on top of the actual robotics involved.

The economics on those robots is that they’ll be used for monitoring and controlling citizens and aliens by the state or employers, rather than be used to serve common folk for their problems including housing

Boston Dynamic robots aren't free. They're far from cheap. Do you think making and sending them out sites would be cheaper than just hiring someone local?

At their current cost? No. But if you could come up with a dependable and repeatable solution that reduced the labor cost of human construction down to about a week of robot construction, then yeah, I think the cost of manufacturing the robots would quickly pay for itself.