Introduction
FANUC is using the term Physical AI for industrial systems shown at IMTS 2026. Its program combines robots, CNC equipment, vision, force sensing and software inside production cells.
Many physical tasks can be automated with fixed or collaborative arms when the workspace is structured. That makes the event useful for comparing humanoids with established industrial hardware.
Information verified from official sources available as of September 14, 2026.
Physical AI is appearing inside existing automation
FANUC's examples start with equipment already common on factory floors. AI is added around perception, process adjustment and robot decision support.
The cell already has known safety zones, fixtures and cycle targets so production results can be measured against an existing baseline.
Force and vision data can close local loops
Industrial robots often need to react to part position, insertion force or tool condition. Combining vision and force data can reduce reliance on perfectly fixed geometry.
Teams should report scrap rate, cycle time, human touches and recovery rate before and after the AI component is added.
Different robot bodies fit different jobs
A humanoid can move between workstations and use tools designed for people. A robot cell can reach high repeatability for one constrained process.
Factories can compare task economics across both approaches instead of treating every Physical AI job as a humanoid job.
Limitations and missing information
- FANUC's event material describes its own products and planned exhibits.
- Trade-show examples should be separated from long-run factory performance.
Conclusion
IMTS 2026 shows Physical AI entering conventional automation where cycle time and quality can be measured against existing cells.
Sources and methodology
Share this article
Share the current TechniaHQRobot article page.
Continue reading
Open the latest robotics reporting, Physical AI analysis and hardware notes.