That's the claim that I'd like to see some data to believe: are, in fact, paid cloud services more long-lasting than free or freemium services? My guess would be no, that a large number of them also go bankrupt, are acquired and shut down, pivot to other lines of work, or otherwise fail to stick around.
I'm with you on your guess. There are hundreds of reasons why a paid service may stop to exist. Still I would argue that running a paid service gives you more options compared to free/freemium: if you balance right you should not go bankrupt, it's cheaper to run (less users -> cheaper hosting and less support), better position to fend off acquisition or during acquisition-negotiations, easier to bootstrap.
