Thing with prefab is it's been tried many times before, and hasn't succeeded. Why? Most companies take traditional construction methods, and simply move them into a factory where there's cheaper labor. But then the assembly process and on-site finish work is laborious and expensive (in this case, 70%+ of costs). For prefab to succeed, there needs to be a ground up redesign of the entire building process, from design and engineering to manufacturing and assembly, with software at the core to enable a system that's scalable and allows for different designs across numerous projects, and with the actual building to be made of parts that truly click together like legos. YC F3 company Cover (i'm a founder) is doing this.
I've been involved in modular buildings in the past and one area that seems to end up informing design choices (in a negative way) is the size of elements that can be readily transported from a central facility to a site. Having bays of modules related to wagon width often, and regular joints/seams seems to end up driving basic ideas like wall loactions and floorplate thicknesses. Have you evolved much in this aspect?
I thought the biggest issue with prefab was lack of demand largely from bad reputation owing to aesthetic or safety concerns - trailers are considered hideous and are often dangerous given lack of sufficient storm anchoring. They were stuck in the role not of a legitimate alternative like siding instead of plaster but 'something to be replaced with a real building as soon as feasible'.
Some of that may be human senses of aesthetics and a sense of uncanny valley with repeats. Just look at video game level generation and the moves to 3D. Preassembled components added more complexity and sense of structure but insufficient amounts of them made the area feel samey and copy pasted.
I'm curious as to how software could be the core instead of merely an optional but useful tool akin to typical usage of CAD for housing schematics.
