Inbolt Real-Time 3D Vision & AI Robot Guidance

Inbolt is a Paris-based company that gives industrial robots real-time 3D vision and AI so they can locate parts and adjust their paths on the fly. It addresses the rigidity of conventional robot cells, where parts must be fixtured precisely and small deviations cause stoppages, and has teams in France, the United States and Japan.

A robot-mounted 3D camera and a generalist vision model trained on CAD digital twins identify each part's position and orientation, and the guidance layer streams corrected motion to the robot controller in under 80 ms. No datasets, labelling or reprogramming are needed. The platform covers vision, CAD-based robot programming and real-time adaptive control, and works with FANUC, ABB, KUKA, Yaskawa, Universal Robots and Comau. Uses include bin picking, depalletizing, de-racking, fastening and dispensing for customers such as Stellantis, Ford, thyssenkrupp and Beko.

• Inbolt Vision Model: generalist 3D model trained on CAD digital twins.
• Real-time guidance: perception-to-motion correction in under 80 ms.
• Robot Programming: CAD-based trajectory generation and validation.
• Robot Control: adaptive motion control layer streaming to the controller.
• Multi-brand support: FANUC, ABB, KUKA, Yaskawa, Universal Robots and Comau.
• Use cases: bin picking, depalletizing, fastening and dispensing.

An automotive assembly plant in Vietnam can automate screw fastening on moving lines. A home appliance factory can let robots pick parts from racks without precise fixtures. A metal parts supplier can add bin picking to feed presses or machining centres.

The company cites training on CAD models in about 5 minutes, full deployment in 24 hours and payback within 6 months. Real-time correction lets robots work on moving lines and with imprecise fixtures, reducing downtime and rejects. One platform across major robot brands simplifies standardisation in multi-site plants.

Automotive and appliance makers want flexible lines that cope with part variation and frequent model changes. Vision-guided robotics by default, rather than costly precision fixtures, is becoming a standard, and the company's move into Japan brings it closer to Asian automotive supply chains.

Paris-based; raised a $17 million Series A in 2024 and an €11 million round led by Shift4Good in 2026. Opened teams in Detroit and Tokyo in 2025. Ranked third fastest-growing startup by Sifted in 2026; reports 40 million robot cycles powered in 2025.