The first and obvious thing he’s good at is vision. he didn’t hop aboard a bandwagon already going full tilt down the road (like Tim Cook taking over from Jobs) he actually as far as I can tell provided the direction from the ground up to get the concept off the ground and see it through to functional, profitable venture. Twice. Three times, if you count the high end luxury EV (Model S and X) and separately, the high volume mass-produced car (Model 3 and Y). I suppose we should partly include the whole Paypal scheme that he helped in…
So many ventures fail because the technical idea guy is not a manager. A simple, small, one or 10-man startup where everyone knows each other, does not translate into a 1,000-plus employee company very easily. So many companies fail to make the transition, lose the idea guy and flounder, or he stays on too long ineptly and causes the company to fail to adapt to its larger size.
So as others have been trying to articulate, this is the first thing. He had the resources and know-how to put together the right team, and motivate them properly. Then to transition to a much larger organization.
As demonstrated in the production hell he described in getting the Model 3 production initial ramp, he provided the motivation and direction (if maybe not the technical expertise) to ensure the process eventually fixed itself, and the willingness to change direction quickly when circumstances were needed.
SpaceX has demonstrated another trait - where the Boeing types seem to have sent their engineers off to a nice tidy office to churn out designs until the decide as a committee “this design will work” Elon seems to have adopted the concept “build something, see if it actually works, and whether it does or not, take what you learned.” The multiple Starship tests provide spectacular examples of testing and failing. he failed spectacularly to re-capture a lot of his early commercial rockets, until nowadays it is rare that one gets away. I presume this concept too is Elon’s idea.
But any technology project today - iPhone, Full Self Drive software, microprocessor chips, toy drones, Windows or MS Office, and especially not electric cars or orbital rockets - none of these are simple enough to be done by a technical group of a dozen people sitting in a converted loft with video games and a pinball machine in to corner to let off steam. There is rarely even one person who is essential to the technology 9as opposed to the management drive).