The robot pitch used to come with an asterisk: impressive on a demo slab, still slower than a good crew on a real deck. That asterisk is coming off. In 2026 the machines that tie rebar and lay brick are posting numbers that beat skilled labor on the jobs they’re suited for, and contractors are putting them on production work instead of pilots.
Start with rebar. TyBot, from Advanced Construction Robotics, crawls a mat on a gantry, uses vision to find each intersection, and ties it at 300 to 450 per hour. A skilled ironworker ties 40 to 80. On a 200,000-square-foot industrial mat that’s the difference between four to six days and eight to twelve, and it’s the kind of repetitive, back-breaking work crews are happy to hand off.
Bricklaying gets a digital-twin brain
Bricklaying is harder to automate because the wall is never quite what the drawings said. Buildroid’s answer is to model the whole sequence first. The startup plans its first U.S. projects in 2026, running Nvidia Omniverse-powered digital twins so multiple robotic masons can work a full trade sequence off a coordinated BIM model rather than improvising course by course. Whether the field tolerances hold up is the open question, and the honest read is that early jobs will tell us more than any spec sheet.
The quieter win is capture
The least flashy robots may matter most. Firms like Hensel Phelps now run Boston Dynamics’ Spot with a mounted laser scanner on complex jobs, walking a repeatable route to log progress and flag deviations without pulling an engineer off other work. Drones and crawler cameras do the same for facades and confined spaces. None of it replaces a crew. It replaces the hours a crew loses to measuring, documenting, and redoing.
Deployments are still small in absolute terms, growing around a third a year. But the trend line has turned from novelty to procurement, and the projects most exposed are the heavy-civil decks and large industrial pours like the Samsung Taylor fab, where the reps are high and the tolerances are unforgiving.