It's both the desire for central control and the Rule of Cool, working in tandem. The customers, the institutions like universities, companies, agencies, campus maintenance teams, etc. like the centralized monitoring and control the Internet of Things potentially offers, yes, and they really like the short-term cost efficiency of using the existing Internet as the data transfer infrastructure. The long term risk doesn't usually get priced in, as you note.
But don't underestimate the degree to which the Rule of Cool is at work among the IT designers, programmers, and engineers who are creating the IoT. It was this same thinking that drove a lot of the rise of the Internet in the 1990s and 2000s, both the useful aspects of it and the huge built-in bad issues derive in considerable degree from the culture that dominate IT work in that period.
'Open, not closed', was one common trope of that culture. A lot of IT people had an idealistic aversion to the very idea of barriers and boundaries within the system, or any kind of coercive limitations on its operation.
Another one was 'information wants to be free', which is semantic nonsense but embodies an ideology common among IT personnel in the rise of the Internet. Jaron Lanier has documented how this ideology, perversely, led to the rise of centralized Internet monopolies like Google and Facebook, which function as surveillance organizations on their own users.
A lot of engineers live and breath the Rule of Cool, if their circumstances are such that they can get away with it. The cliché of the creative engineer or programmer stifled by an accountant or bureaucrat contains a germ of truth, and it's also true that sometimes the accountant is right, and his presence is necessary to keep the creative engineer from making something hugely Cool but irrelevant to the customer or a liability to the organization.
A lot of the technical people working to create the IoT are motived less by profit (though of course they hope to make money) than by the sheer fun of it. It's a Cool idea, from the technical POV. Whether it's a good idea is a separate question entirely.