FANUC announced that the system, developed with Google using Gemini Enterprise, reads an engineering drawing and generates welding parameters and robot motion programs. The operator can execute those outputs as-is or fine-tune them before welding.
Mechanism: The system captures a drawing, interprets material and part shape, and generates welding current, voltage, and robot motion instructions; operators may run or fine-tune the outputs.
Why now: FANUC cites shortages of skilled welders who can optimize parameters and teach robot motions, and says these shortages contribute to longer manufacturing lead times. The company plans a demonstration at the International Welding Show beginning September 16 and shipments by the end of December 2026.
First implication: The announced workflow shifts part of robot programming from manual teaching toward review of generated task instructions, but does not establish that human oversight can be removed.
Evidence boundary: The source is FANUC’s announcement; it provides no independent validation, quantified productivity results, error rates, or customer deployment evidence.
First watch condition: Check whether the scheduled shipments lead to customer deployments with measured setup-time, weld-quality, and correction-rate results.