First, if academics wanted to work on topics mandated by somebody else, they would go work in industry for that somebody, and earn much more money than they earn right now.
Second, most academic scientists do not do anything relevant to Musk's companies. Do you expect a chemist to pivot to self-driving cars? Or a pure mathematician to whatever X.ai is doing?
The only thing this will lead to is a destruction of American capacity to carry out independent scientific research.
Very few academics become principle investigators. Most every academic who's not a PI is working on something for that PI.
> inherently trainee positions
Do we live on the same planet? I understand that the point of being an academic is to always be learning, but there's no place on earth I know of that thinks of someone with a PhD as a trainee.
> the main options are becoming a PI or choosing a teaching-focused position. There are some staff scientists and similar, but such positions are rarer than tenured professors.
Implying that one gets a choice is bold. My understanding is that there's a job for about 1 in 10 postdocs in academia these days.
Receiving a doctorate does not mean that you have finished your training. Some countries have habilitations or higher doctorates, which can be understood as more formal versions of postdoctoral training. Medical doctors are expected to specialize and receive more training as residents. Other fields have similar arrangements, some more and others less formal. If a full career is 50 years and the job requires a high degree of specialization, it can make sense to use the first ~15 years for training.
A postdoc is primarily a career advancement position rather than something where you are expected to contribute full time. Such positions are also pretty rare. There are something like 70k postdocs in the US, vs. almost 190k tenured or tenure-track full-time faculty in research universities.