My gut tells me that fusion power is better analogized to a space elevator than reusable rockets.
But you're suggesting what I already alluded to as a reason for doubting that economic fusion power will ever happen - he's the type of person who would be pursuing fusion if it seemed feasible. It's not that he's the only one, or that he knows about nuclear physics, but it's the sort of risk he would plausibly take if he didn't know of a good reason why it can't be done. Also, of course, you need power for Mars bases.
I'm not a huge SpaceX watcher, but I suspect it's harder to build electric cars than rockets. Yes, rockets are crazy expensive and time consuming but the fundamental principles are relatively understood.
Tesla's continual hitting of projected targets is much more impressive to me.
You're generally right though: the maturity, risk, and timelines of fusion are much greater than either of these. Better suited to long-term gov't/military spending.
But the general principle applies. Everything is measurable and can be modeled with a predictive distribution. The more unknowns, the more variable the distribution is.
