His speciality seems to be leading companies that solve tangible measurable hardware problems like "cheaply launch orbital rockets" and "make commercially viable long range fast electric car" (making cheap underground car tunnels is impossible so it doesnt really matter how good a hardware guy he is).

I don't know why you'd think being good at a domain would transfer to other very different domains. There's no reason to think an amazing singer would be a good mathematician, why should a hardware production line maker have good taste in truck aesthetics?