Only slightly tongue in cheek: Imagine the focus you could achieve with dedicated personnel taking care of all of the operations concerns for your devs...
If you know good ops people, you have additional options. Hiring ops people and hoping they can out-value a cloud provider is a risk. Good ops people can cost as much as good developers.
If you're a larger business, you can probably absorb that risk if it goes wrong. If you're a startup, it could be catastrophic.
However, I also know larger businesses that just simply do not have good ops teams and a cloud provider would outperform all of them.
Almost every argument that you are making here could be exactly applied to developers and contractors. Hope that they can out-value a contractor. They're expensive. Some larger businesses have terrible dev teams and a team of contractors could out-perform them all.
> Good ops people can cost as much as good developers.
And this is a surprise? They bring a ton of domain specific expertise, good automation experience, and they lift the burden of managing your systems from your developers, so they can work on features and not scaling.
If you can afford to put good people in every position, by all means go for it, but that's not been how startups and small businesses hire, or else they wouldn't have to use the phrase "we wear a bunch of different hats here at St4rtUp". The apparent trend has been to overload the developer with additional responsibility instead of, e.g., making ops people build product features.