Real robots. Real work. Made here.

American manufacturing did not lose automation because the work stopped mattering. It lost it because the machines, the parts, and the expertise moved elsewhere. Rebuilding that capacity is not a matter of nostalgia or tariffs. It requires being genuinely better at the thing that comes next — and the thing that comes next is Physical AI.

Physical AI — embodied AI, in academic circles — is artificial intelligence that perceives and acts in the physical world through a robot body rather than on a screen. The foundation models already exist. What they lack is real-world demonstration data captured on real robot hardware, and that is the gap Yuri fills. A robot that can be taught a task by a person, and then perform it autonomously, is a fundamentally different proposition from a machine that has to be programmed for one job and re-tooled for the next.

Bringing that home means doing several unglamorous things at once. It means designing and assembling robots here, and sourcing the parts here — more than 65% by price today, and rising as the domestic supply chain rebuilds. It means writing the software ourselves rather than leaving integration to the customer. It means training the people who use these machines, and supporting them from the same building where the robot was built. And it means putting real hardware into the hands of the labs inventing this field, so the research happens on American robots.

That is why we start with research platforms and do not intend to stop there. Every system we place teaches us which tasks robots can perform reliably. The wheeled Yuri Mobile configuration is where that work moves off the bench and into real rooms — and toward robots bought for the work they do.