Tesla’s factory automation advantage
Elon Musk asserts the Tesla Model 3 will, at some point, be assembled by robots and never touched by a human (Update it may be the Model Y). Dave L recently claimed that any auto manufacturer could do this. I believe it’s unlikely other auto manufacturers will do automate to this degree and here is why.
Some auto manufacturing is already automated, such as welding, gluing, and painting. However, final assembly, installing interior, wiring etc. is done manually. The model 3 is designed to have general assembly be done by robots. The large roof will allow robots easy access to the cabin. The components such as fascias and harnesses are designed to easily pop in place by a robot. Plus EVs are inherently simpler vehicles. This type of integrated approach, is not on the radar for most automakers. While big auto will continue to make incremental improvements with suppliers, such as Kuka and Fanuc, Tesla is betting the farm on this approach that is highly integrated with the design of the car.
Peter Hochholding, Teslas VP of vehicle production and former Audi executive states:
I think Tesla is the first automotive company that is thinking that way. I’ve never seen that before. I believe if you have a machine that makes a machine — if you have a smart machine that makes a machine — you can create more throughput, more quality.
Another example of Tesla’s speed increases, Greg Reichow, former Tesla VP, stated
Some of the robots moved at such high speeds that their arms needed to be built from carbon fiber instead of steel.
The goal is to lower labor costs and increase the speed of manufacturing. An increase in speed will allow Tesla to increase the number of cars while building less factories than competitors. The calculus for existing auto manufacturers that already have factories and workers doesn’t make as much sense as it does for Tesla. It’s a lot easier to justify extreme investment in automation for Tesla because they won’t be laying people off, or selling off factories. They will be getting more cars to market faster by investing in automation rather than raising more money (and diluting their stock) to build more factories. It’s the classic conflict of interest faced by incumbents as outlined in the Innovator’s dilemma.
Another reason to believe Tesla will dominate factory automation is their lead in autopilot. I suspect the same technology in autopilot, presumably deep convolutional neural nets, will also be the enabling technology for the alien dreadnought (fully automated factory). Think of it this way, workers will likely assemble the Model 3 manually by guiding robots equipped with cameras and sonar while recording sensor data and the corresponding robotic movements. Over time they will collect training data about how to assemble the car, just as they collect driving data today. Gradually they will be able to train neural nets to assemble the cars without human intervention.
The details of this last part are educated speculation. Please comment if you know any technical details about the alien dreadnought.