I hear this debate repeated often, and I think there's another important factor. It took me some time to figure out how to explain it, and the best I came up wi

I hear this debate repeated often, and I think there's another important factor. It took me some time to figure out how to explain it, and the best I came up with was this: It is extremely difficult to bootstrap from zero to baseline competence, in general, and especially in an existing organization. In particular, there is a limit to paying for competence, and paying more money doesn't automatically get you more competence, which is especially perilous if your organization lacks the competence to judge competence. In the limit case, this gets you the Big N consultancies like PWC or EY. It's entirely reasonable to hire PWC or EY to run your accounting or compliance. Hiring PWC or EY to run your software development lifecycle is almost guaranteed doom, and there is no shortage of stories on this site to support that. In comparison, if you're one of these organizations, who don't yet have baseline competence in technology, then what the public cloud is selling is nothing short of magical: You pay money, and, in return, you receive a baseline set of tools, which all do more or less what they say they will do. If no amount of money would let you bootstrap this competence internally, you'd be much more willing to pay a premium for it. As an anecdote, my much younger self worked in mid-sized tech team in a large household brand in a legacy industry. We were building out a web product that, for product reasons, had surprisingly high uptime and scalability requirements, relative to legacy industry standards. We leaned heavily on public cloud and CDNs. We used a lot of S3 and SQS, which allowed us to build systems with strong reliability characteristics, despite none of us having that background at the time.