[notahacker]

The alternative to getting on the tenure track is usually getting a commercial role, probably a relatively well paid one. There is literally no worse reason to start a startup than "trying to get tenure is getting more difficult". Because if your end goal is to study phenomenon that particularly fascinate you at your own pace in a collegiate environment, then starting a startup is a big step in the opposite direction. And if it's the difficulty or possibility of failure that's putting you off pursuing a career in academia...

Sure, many PhDs are eminently capable of starting a startup and some will be wildly successful. But that's more likely to be because they have a burning desire to commercialise their research or a great founding team sees their skillset as the missing link than because they couldn't find an alternative