Advanced Metal Research

Factories that buy industrial robots from established makers face hidden friction well after purchase. Motion functions, process packages and programming seats are licensed piecemeal on robots the buyer already owns. Each maker uses closed, proprietary languages, so programs do not transfer between brands and outside software, including AI tools, gets only limited access. Service is often slow, opaque or expensive, and when a production line goes down the wait itself becomes the cost. Buyers stick with these vendors mainly because the hardware is reliable, not because the surrounding software and service experience is good.

Advanced Metal Research makes the Rosie 1400, a six-axis industrial arm with 1,400 mm reach, and War Machine, a production cell built around it with a second bay for part loading. The controller, RosieOS, is open source under Apache 2.0 with every function unlocked, and can be programmed from the pendant, in code, or by an AI agent through a documented control API and open sensor data. Paid, optional skills sit on top of this open stack, starting with a welding intelligence package that includes seam tracking within 0.20 mm, weld-pool imaging, arc-sound fault detection, and adaptive control of path, torch and heat during welding. AMR also runs a metallurgical lab that sections and tests welds for hardness and microstructure to ISO 13919-1, feeding that data back to train the system's weld physics simulator and AI world model. Every Rosie 1400 ships with the arm, controller, drive cabinet, cabling, pendant and on-site commissioning.

The site cites the American Welding Society estimate that the US needs 320,500 new welding professionals by 2029, framing welding automation as an urgent labor gap for AMR's first paid skill.