Americans have a very 1980s idea of manufacturing (and China in general) in that there aren't actually that many humans being used in Chinese factories let alone the American ones some of them want to build here. There's even a concept of, "Dark Factories" in China which are 100% automated factories that operate in the dark. The only jobs that will come from bringing manufacturing back to the states will be in automation, robotics, AI, and roles to support those things.
Given the all the minimum wage staffing at most distribution centers these days despite all this off the shelf robotics technology seemingly available on order and already proven, makes me thing the american worker is cheaper than we might suspect compared to building out these dark amazon warehouses.
A business I work with has a factory in China that produces their devices. They absolutely do most of the assembly manually, as many of their sister factories do.
Robotics automation is a tradeoff to gain efficiency at the expense of flexibility, with a large upfront cost.
Well, even a better argument to bring those factories to the US. Why not develop the know-how on manufacturing and improve automation in the US rather have China lead there.
Because automation is expensive. It pays off in volume. A skilled human can often build a single widget faster than an engineer can write the automation for the robots (because a skilled human will see parts that don't fit and "file to fit" while the robot demands more effort to double check all that). When you only need 10, the program is faster to write, but you still need to pay for the robots and they are expensive (often $million each, while the human is only a few thousand for his time)
Of course there are a lot variables in the above. As time goes one automation gets better. You can buy cheap robots for some common operations, and a good engineer with good CAD can run various automated analysis to ensure fit and then export to the robot and build even a single part cheaper than the human - amortizing the cost of the robot over thousands of different single parts made this way. However as the widget gets more complex you reach the point where humans are needed. In some cases you just have humans to take the parts off of one machine and put them into the next, but it is still humans. We can automate even that, but often the robot to do that would cost more than a human for 10 years.
