The most expensive part of a factory is often the human hand. It’s the part that picks a delicate component from a bin of tangled parts, the part that packs a finished good without crushing it, the part that assembles something with a dozen tiny, fiddly pieces. 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 not in raw speed but in a 99%+ success rate on tasks that currently require a person [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 simulated and real-world 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,a part that’s slightly askew in the bin, a box that’s not perfectly aligned on the conveyor. The company claims its algorithms can theoretically interface with any robot hardware, which sidesteps the capital-intensive work of building new mechanical systems and lets them focus on the software intelligence [Usine Nouvelle]. This is a pure-play on the unit economics of dexterity: if they can drive down the cost and time of training a robot to handle complex manipulation, they can open up automation for a wider swath of small-batch, high-mix manufacturing.
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, an experience that involves coordinating complex, high-stakes logistics under immense pressure [Reuters, Lesfrancais.press]. CTO Roch Molléro previously worked as an AI consultant, focusing on audiovisual solutions, which suggests a background in translating complex sensory data into actionable outputs [Crunchbase]. It’s a pairing that speaks to the company’s dual challenge: navigating the rigid, safety-first world of industrial sites while pushing the frontier of AI training pipelines.
The competitive landscape and the path to proof
Gobano enters a field with established heavyweights and nimble specialists. The competitive set includes everything from the iconic research of Boston Dynamics to focused automation players like Genesis AI and Linkerbot. Gobano’s differentiator is its narrow focus on dexterous manipulation as a software layer, not as a full-stack hardware play. This allows for faster iteration and lower upfront cost for potential customers, but it also introduces dependencies on hardware partners and places the entire burden of performance on the AI model’s robustness.
The €3 million pre-seed round, led by Axeleo Capital with participation from Bpifrance Digital Venture, Polytechnique Ventures, and others, 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]. The next twelve months will be about moving from the theoretical “can interface with any robot” to a demonstrated “has interfaced with this robot, at this customer’s site, for this task, and it works.”
| Competitor | Primary Focus | Gobano's Differentiator |
|---|---|---|
| Boston Dynamics | Advanced mobility & manipulation (full-stack) | 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 |
A back-of-the-envelope calculation shows the scale of the bet. If a human worker costs ~€50,000 annually in a high-wage European market, and a Gobano-equipped robotic cell (hardware plus software subscription) can replace two such workers per shift across three shifts, the payback period drops into the realm of a standard capital expenditure for a manufacturer,provided the robot’s 99% performance claim holds under real factory conditions. The company isn’t just selling a robot arm; it’s selling the delta between the cost of human dexterity and the cost of automated dexterity. To win, Gobano must prove its AI can consistently beat the incumbent,the trained human hand,on both cost and reliability.
Sources
- [Gobano website, retrieved 2026] Gobano Robotics - AI Robotics automating dexterous tasks | https://www.gobano.ai/
- [Vestbee] French robotics firm Gobano Robotics raises €3M to accelerate R&D | https://www.vestbee.com/insights/articles/gobano-robotics-raises-3-m
- [Usine Nouvelle] Gobano Robotics company information | https://www.usinenouvelle.com/
- [Global Industrie] Global Industrie exhibitor page | https://mygi.global-industrie.com/event/gi-2026-salon-exhibition/exhibitor/RXhoaWJpdG9yXzIzMzY4NzY=
- [Reuters] Euro 2016 security chief 'cool and confident' despite militant threat | https://www.reuters.com/article/us-soccer-euro-security-chief-idUSKCN0Y71KN/
- [Lesfrancais.press] « Ziad Khoury: l’enfant du Liban devenu Préfet » | https://lesfrancais.press/ziad-khoury-lenfant-du-liban-devenu-prefet/
- [Crunchbase] Roch Molléro - Founder and CTO @ Gobano Robotics | https://www.crunchbase.com/person/roch-moll%C3%A9ro
- [Welcometothejungle.com, retrieved 2026] Gobano Robotics: photos, vidéos, recrutement | https://www.welcometothejungle.com/en/companies/gobano-robotics/jobs/mechatronics-robotics-engineer_nantes