The most expensive part of a factory is often the human hand. Gobano Robotics, a Paris-based startup, is betting a fresh €3 million that the best way to automate that hand isn’t to build a better robot, but to build a better brain for the robots already available [Vestbee].
Their pitch is straightforward: take a standard robotic arm from a partner, train a bespoke AI model on the specific dexterous task at hand, and deploy it on a factory floor within a few weeks [Gobano website, 2026]. The focus is on the messy, variable tasks that have resisted traditional automation, soft manipulation, bin picking, packing, and assembly, where success is measured in a 99%+ success rate [Usine Nouvelle].
The AI wedge into physical work
Gobano’s approach leans heavily on imitation and reinforcement learning, creating a pipeline where a robot can learn from human demonstrations and then improve through practice [Global Industrie]. The core promise is adaptability. Instead of programming a robot for one rigid motion, the AI model learns to handle the natural variability of physical objects. The company claims its algorithms can theoretically interface with any robot hardware, which sidesteps the capital-intensive work of building new mechanical systems [Usine Nouvelle].
Founders from the frontier of logistics and AI
The founding team brings together an unusual blend of operational scale and technical depth. CEO Ziad Khoury is a former senior French civil servant who served as security chief for Euro 2016 [Reuters, Lesfrancais.press]. CTO Roch Molléro previously worked as an AI consultant, focusing on audiovisual solutions [Crunchbase].
The competitive landscape and the path to proof
Gobano enters a field with established heavyweights and nimble specialists. Gobano’s differentiator is its narrow focus on dexterous manipulation as a software layer, not as a full-stack hardware play. The €3 million pre-seed round provides runway to move from a tested MVP into real-world pilots [Vestbee]. The company is hiring a mechatronics engineer in Nantes, indicating a push to bridge the gap between their AI models and physical integration [Welcometothejungle.com, 2026].
| Competitor | Primary Focus | Gobano's Differentiator |
|---|---|---|
| Boston Dynamics | Advanced mobility & manipulation | Software-only layer for dexterous tasks on partner hardware |
| Genesis AI | Vision-guided robotic automation | Specialization in high-variability, soft manipulation tasks |
| Linkerbot | Collaborative robotics | AI training pipeline for rapid deployment of specific skills |