Built Robotics and xLAB Launch Physical AI Safety Research Pilot on Active Construction Sites

The Daily Dig

Built Robotics and xLAB, the Safe Autonomous Systems Lab at Penn Engineering, have announced a research collaboration focused on advancing physical AI models for construction. Safety is the central objective for both organizations.

The first phase of the pilot will deploy Built Robotics' edge AI model across a fleet of construction survey robots operating on active solar projects. Those robots will collect high-fidelity sensor data in real field conditions, which Built will use to refine its existing AI models and expand them to additional vehicle platforms and construction activities.

Built says its proprietary edge AI model has been refined across active construction sites where it handles personnel detection and human safety monitoring, including large-scale sites with hundreds of workers spread across thousands of acres.

The partnership gives xLAB researchers access to live jobsite data and real-world operational parameters, something Mangharam described as essential to bridging the gap between validation in controlled environments and actual field performance.

Built CEO Noah Ready-Campbell, a Penn Engineering alumnus, cited Dean Vijay Kumar's work on quadcopters and multi-robot coordination at Penn's GRASP Lab as a formative influence when he started the company. He said the mission alignment with xLAB became clear as Built's robot fleet scaled in the field.

TheJobWalk Thoughts

Built has established its AI track record in utility-scale solar, and that is exactly where this pilot is running. For subs and suppliers already working that sector, autonomous survey robots collecting site data are no longer a concept being tested somewhere else. They are being deployed on active solar projects as part of this pilot, and the data gathered will directly shape how these systems develop across the industry.

The expansion intent is stated plainly. Built will use the sensor data from this pilot to extend its AI models to other vehicle platforms and construction activities beyond survey robots. Solar is the starting point, not the finish line.

xLAB specializes in developing provably safe, highly resilient software architectures for autonomous systems, which is a meaningfully different standard than general AI development. As physical AI moves closer to mainstream jobsite use, owners and GCs will want to know how these systems were built and validated before they approve them on their sites. Built investing in that foundation now, rather than solving for safety after the fact, is the kind of detail worth noting when this technology eventually lands in your procurement conversation.