Mainframes are awesome, but they’ve achieved what the hyper scalers aspire to - complete market domination. They finished up in 1991 or so.
The problem is there’s no growth. There are no new customers, no new patterns or meaningful growth areas. It’s a career cul de sac, and when the bean counters at your big bank discover Oracle Cloud ROI is 6% more than you, you get nuked from orbit.
CIOs have a bias to be wimps. It's easier to kick the can down the road.
I work with alot of orgs that have these things, both directly and from a association/conference perspective - it's starting to change because the new CIOs didn't grow up with mainframes, and are finding themselves with no alternatives other than expensive and increasingly ineffective contracts to operate these things. 1990 was 35 years ago, the old guys are dead - the can has been kicked off a cliff.
IBM Endicott is being demolished this month.[1] IBM hadn't been doing much there in the last 20 years, but now it's all being flattened.
This underestimates the switching costs from mainframes. As one CIO told me, it’s basically career suicide. You take a rock solid system that’s been working forever and embodies decades ago of knowledge and replace it with something that while modern and cheaper exposes you to massive risk.