The interesting question about humanoid robots is no longer whether they work. More than 100 OEMs are shipping or close to shipping production units, unit prices have fallen from above $150k in 2022 to roughly $30k for some models today, and Goldman Sachs now expects 1.4 million units shipped in 2035 into a $38B market, up from 100,000 in 2026.
Hardware has stopped being the binding constraint. The question that matters now is who gets these machines doing useful, reliable, insured work inside a real European factory.
The problem: Buying a robot is only the start
Ask a plant manager in Ghent or Dortmund why there is no humanoid on the line, and the answer is almost never that the robots cost too much. Four things stand in the way of early adopters today:
Choice. Which of several hundred tasks on a site is worth automating, and which of dozens of available models can actually do it? Nobody inside the plant is equipped to answer that.
Training. A humanoid arrives as a general-purpose machine and has to be taught a specific job. Most mid-size companies do not have the in-house robotics talent, and specialist integrators quote 4–8 weeks per task at consultancy rates.
Capital. A deployment is an upfront hardware purchase against an unproven return, landing in a capex budget that was set last year and defended by a CFO who has never seen a humanoid in action and who is operating in economically unstable times.
Trust. A humanoid is a networked, sensor-heavy machine sitting in the middle of a company’s operations. Industrials want to know where the data goes and who is accountable for it, and they want that answer to be European.
Pour lire la suite : https://medium.com/extantia-capital/putting-humanoids-to-work-on-europes-factory-floors-a18febbf0757
