It's a way to "commoditize" engineers. You can run on premise or mixed infra better and cheaper, but only if you know what you are doing. This requires experienced guys and doesn't work with new grad hired by big cons and sold ad "cloud experts".
(I do that for people; my AWS using customers consistently end up needing more help)
but somehow that is never a problem.
When everyone is suffering because AWS is having its bi-yearly 8 hour outage, the CTO isn't blamed, bonus all round, and maybe the AWS sales team takes him for an apology lunch
When the CTO is up for 1500 days straight then has a 2 hour downtime when nobody else does, the CTO is blamed, no bonus, and more likely to get fired
IBM is older, and it's incredibly well documented how mainframes are more expensive to run than normal servers.
this is the main advantage of cloud, no one cares if the site/service/app is down as long as it's someone else's fault and responsibility.
If not, you're at the mercy of others.