The difference was mostly a choice, and it's arguably the main "smart" thing Musk was able to do, whereas Holmes dropped the ball on. Musk's first breakout gig, Paypal, was honestly just dumb luck of being in the right place at the right time - online credit card payments were an inevitability, and it was just arbitrage on getting there before the banks got their feet wet. That is - at the time, "everyone" broadly agreed it was going to happen, even banks, but some groups were institutionally slower about setting it up. He happened to be lucky by getting in on a startup right then, when low-hanging fruit like that was waiting to be plucked.
The left him wealthy enough to fund something else as a sort of Angel Investor, and it was at this point that "his choice of what to fund" was a significant act of cleverness. He looked at the world and asked "what are some things that are technically totally viable, but have been completely stagnant because of political problems in getting funding to try them, rather than any technical problems with them actually working? He spent a few years(!) crunching the numbers on them, and doing some research, and concluded that two things PROBABLY fit that bill: electric cars, and reusable rockets.
What he was ideally looking for is something where an industry had a blind spot - where the industry had "tried it, failed, and decided it's a dead idea nobody should bother trying again". But where they'd done this for a really bad reason.
----
These were two things that just needed some billionaire to finally take them seriously - they weren't something some "rogue scientist" could just do in their garage; you genuinely needed major capital investment for said rogue scientists to be able to build test versions and refine their ideas into something that worked.
Electric cars were wide open because all the basic tech was there, but the existing car companies weren't even interested in trying due to cultural reasons. The biggest thing is most of the critical advances in batteries had already been driven by hundreds of billions of R&D dollars in the consumer electronics and computer industry - they'd been desperately pushing battery tech forward, and it'd become orders of magnitude better than it was when the auto industry made a consensus decision that electric cars weren't viable (back in the 80s). In the 80s they had a range of maybe 15 miles, and struggled to get up hills. Auto execs just didn't understand HOW much better batteries had gotten in a mere 10 years (in fact they probably assumed batteries weren't even really improving, and had reached their physical/scientific limits).
Reusable rockets had been scrapped because of the space shuttle's failure; it'd been intended to be completely reusable, and they couldn't pull it off because ... reasons. The consensus was that those "devil in the details" issues would always come up. On closer investigation, though - most of what went wrong is the same procurement malpractice we're currently having with stuff like the F-35 fighter jets, and not something that had any real bearing on the actual engineered work itself.
Similar thing with Jobs - computer GUIs existed, but the only companies that had developed them had been sitting on them and refusing to deploy them as a mass-market product.
----
The key difference with Holmes is the fundamental tech keystone (doing medical tests via micro-samples of blood) just wasn't viable. Some biotech researchers tried to sound the alarm about that, but once they were too far into the process of running the company, everyone had already laid down their financial bets and was afraid to pull out.
What happened to them is very similar to what would have happened to SpaceX if it turned out that they were wrong about reusable rockets (for example, if SpaceX had tried to be a company in 1995, before modern computing was able to do all the differential-equation wizardry that lets their rockets land). Or if Tesla had tried to make an electric car company in 1988, and bitterly realized "oh shit, there's no humanly possible way we can make a vehicle with more than 30 miles of range."
You have to take a bet that everyone else is betting against ... and then also be RIGHT about that wild, contrarian bet.
----
Choosing that bet was the smart thing Musk did. It's possible to make wickedly smart bets like that, and still be a mediocre manager, but win because the bet itself was just such a bullseye.
One reason for this is that sufficiently good bets can often attract great managers; both Tim Cook, and Gwynne Shotwell, especially, may be great examples of people who've come in under the auspices of a "big idea guy", but have been able to make their companies excel because they replace the poor managerial skills of the idea guy with their own.
But yeah; a guy like Musk's time is best spent not micromanaging their companies, but is best spent strategizing on big picture stuff (like doing the number-crunching to reveal that SpaceX could enable Starlink).