It was Optimus’ big day. After several months of heating up public interest in himself, Elon Musk presented two prototypes of his mysterious humanoid robot on the evening of September 30th. The first one took the Tesla AI Day stage and waved to the audience. A clip released at the event also shows him delivering a package to a Tesla employee or watering plants.
Another prototype, unveiled by the Tesla team, is more complex, with much less visible cables, but cannot yet move or stand. What Tesla teams have been able to develop in such a short time is impressive. Humanoid robotics is indeed an extremely complex discipline.
The bet that Elon Musk makes in it is not a simple whim. First, Tesla has an extremely advanced knowledge of computer vision, which is a key element of its semi-autonomous vehicles. This gives him a head start in robotics, as androids also need to properly analyze their environment in order to move around without hurting anyone, hurting anything…or hurting themselves.
Limited offer. 2 months for 1€ no obligation
Tesla, which has giant factories and wants to automate the production of its electric vehicles even more, will be Optimus’ first customer, giving the latter commercial outlets even before it’s released. Finally, the humanoid form chosen by Elon Musk has undeniable advantages, as it gives robots greater versatility: theoretically, they can evolve in a variety of environments (uneven terrain, warehouses with stairs, etc.).
Awesome but inefficient robots
Betting everything on this humanoid form, however, is risky. These designs are indeed much more difficult to manufacture than geometrically shaped robots. Thus, the humanoid robot Ameca, which made a splash at CES 2022, consists of several thousand standardized components and more than 1,000 custom-made modules. “The challenge is that the human body relies on many axes: wrists, elbows, ankles, knees… So we have to replicate it all with mechanical axes and motorized systems,” explained Cyril Cabbara, CEO of Shark Robotics, last May at L ‘Express.
If Elon Musk pointed out during Tesla AI Day that the Optimus should end up costing “less than $20,000”, then we’re still very far from it. All competing humanoid robots are currently hand-built and cost a small fortune. The problem is that their abilities aren’t much better in all areas. For example, simply standing up requires energy from these metal bipeds, which seriously affects their autonomy. They also don’t compete with tracked or wheeled robots when it comes to moving quickly or carrying heavy loads.
Therefore, for humanoid robots to become useful, huge technological advances and drastic cost reductions must be made. However, Tesla teams already have a lot of work to do to implement truly autonomous driving in their vehicles. Scattering is a bold bet. It remains to be seen whether Elon Musk will refute the alarming predictions, as he did with Tesla and SpaceX, or whether Optimus will become part of his list of failures, like SolarCity.
Chronicle of Christoph Donner
Detours from France