A masonry robot is only as good as the plan it’s handed. That’s the bet behind Buildroid, the startup now putting bricklaying robots on U.S. jobsites after emerging from stealth last fall with $2 million in pre-seed funding. The pitch isn’t faster hardware. It’s planning the whole wall before a single block gets placed.
Founders Slava Solonitsyn and Anton Glance built the system around a digital twin. A plug-in for Autodesk Revit converts a building model into OpenUSD format plus a YAML sequencing file, then Buildroid runs the full trade sequence inside an Nvidia Omniverse simulation, pushing the model from LOD300 detail up to LOD400 or LOD500 before any robot ships to site.
How the masonry robots work
On the jobsite, the company runs two block-laying machines and an autonomous mobile robot that handles material. The placing units lift blocks up to 40 kilograms and build walls up to 4 meters wide and 3 meters tall. Coordinating multiple robots across a real trade sequence is the hard part, and it’s where the simulation pays off: the digital twin tests the choreography so the hardware isn’t improvising in the field.
A shared-savings model and a $13B target
Buildroid plans commercial work in the first quarter of 2026 with general contractors, starting on blockwork and partition-wall installation. That’s a $13 billion slice of construction, and the company isn’t selling robots by the hour. It charges on a shared-savings model, taking 50% of the net efficiency gains and committing to specific performance metrics.
That structure is the interesting wrinkle. Masonry and bricklaying robots have a long demo-reel history and a short jobsite one, and a shared-savings deal puts Buildroid’s revenue on the same line as its results. If the robots underperform, the company doesn’t get paid. Compare that to the equipment side, where August Robotics and Caterpillar’s autonomous machines are chasing the same physical-AI moment from the earthmoving end.
Blockwork is unglamorous, repetitive, and short on labor. If a robot can lay it reliably and the developer only pays for savings, the adoption case mostly writes itself. The question is whether the field matches the simulation. ENR has the rollout details.