Residential construction has raised productivity 0.45% a year since 1993

Put the $11 billion a year that labour shortages cost homebuilders next to the shortfall itself: 11 billion divided by 1.2 million homes is about $9,200 for every home the country does not have, each year ^[5][1][1]. The $9,200 is a builder's operating cost, not the value of the missing houses. It is also the size of the market an error-prevention vendor sells into.

Robert Dietz's productivity number does more damage than his dollar one. A 16% gain over the 33 years since 1993 compounds to roughly 0.45% a year, and the wider economy's better-than-50% works out near 1.2%, so residential building has improved at a bit over a third of the national rate ^[4][2]. Nearly 300,000 construction jobs stood open at the end of 2025, about 41% of the workers the NAHB says the industry has to attract in a single year ^[2][3][3].

Vineet Kamat, who studies construction robotics at the University of Michigan, put the screening test in commercial terms. "The key question is not whether a robot can work in a pilot," Kamat said, "but whether it can do so reliably across varied conditions with a positive business case." ^[10]

Tessa Lau, who founded Dusty Robotics, said factories are relatively easy to automate because a robot repeats the same task in a fixed environment, while on a site the robot's workplace is being transformed around it ^[13]. Patrick Murphy, chief investment officer at Coastal Construction, said the lack of uniformity across designs, soil conditions, architects, subcontractors, materials and owner requirements makes repetitive automation difficult ^[12]. On the humanoid version, Murphy said: "The biggest misconception is people imagining humanoid robots walking around job sites hanging drywall and painting. We're years out before we're in that world." ^[11]

Dusty's product is narrow by design. The FieldPrinter takes digital building plans and prints them onto the floor, marking walls, plumbing, electrical and mechanical locations along with text, images and QR codes, so several trades read the same instruction ^[14]. Lau built it after a plumber installed the wrong fixtures in her two bathrooms, a mistake she found only after the tubs had been tiled; workers chipped out the tile, removed the fixtures, reinstalled them and retiled ^[15]. "That's an information mistake," Lau said ^[16]. "If you're not thinking about it very deeply, the shallow take is that labor replacement is the solution," she said ^[19].

The reallocation case is strong on where gains are measurable and weak on what they reach. Dietz said "There's no single, simple, scalable solution," pointing to labour, regulatory costs, land availability, supply chains and financing ^[7], and a printer on the slab does not touch land or a construction loan. CNBC's assessment is that the gains so far are real on individual projects and too narrow in scale to materially reduce the national shortage ^[17]. Against that, the two months labour shortages add to the average build is a carry cost ^[5]. A builder who takes that back on every house books the saving whether or not the national series moves off 0.45% a year. Movement in that series is the reading that would show these tools reaching further than this evidence supports ^[2]. Asked whether robotics is already improving homebuilding productivity, Dietz said, "It's too early." ^[6]