Buildroid AI pays half of the efficiency gains of its robot masons

The San Francisco start-up rehearses the projects in Omniverse digital twins before sending any robots, and invoices by sharing the savings made.

Founded by Slava Solonitsyn and Anton Glance, which emerged from secrecy in November 2025 with $2 million in pre-seed funding led by investor Tim Draper, Buildroid AI is based on an unusual principle in construction site robotics: simulate first. The platform builds a digital twin of the construction site on NVIDIA Omniverse, tests the complete sequence of tasks and deploys the hardware only afterwards. It claims compatibility with more than forty types of robots, coordinated in integrated workflows via computers installed at the edge that serve as a gateway.

The catalogue is based on a block laying robot, complemented by a compact variant for residential plans, an autonomous mobile robot for material delivery and plastering machines that are still under development.

The business model is also a statement: robotics is sold as a service, and the company is paid half of the net efficiency gains observed, by committing to precise performance indicators. It is targeting the blockage and partition niche, estimated at $13 billion in a global construction market of $17,000 billion, where it crosses Construction Robotics, Fastbrick Robotics and Monumental.