It's sort of Tesla's core thesis that they want to attain the world's lowest manufacturing costs, by doing novel things with robotics and AI that automate away humans.

Everyone gets this backwards. Tesla isn't abandoning their low-margin auto manufacturing in order to switch to high-margin robotics tech; they want to capture the margin of robotics for themselves first. If their manufacturing half fails, that means their robotics half also failed at its primary task.

Techworld has an inductive bias towards expecting successful products to be luxury lifestyle products, because so much of past tech successes came from that. Most likely the first successful AI robotics aren't going to be iPhone-like consumer luxuries, laundry-folding novelties for bored techies. They'll rather continue the trend of robotics to focus on industrial processes in controlled settings, capturing value by lowering the costs of doing banal things.

Chinese cars are effectively excluded from the US market mainly to keep Tesla afloat.

Which begs a question: Is it really possible to compete with China on consumer products? The price and cost of labor advantage, indeed state sponsorship and support, would suggest that it's not.

Musk basically lost to China competitition, but can't assume it.