Except that nobody watching benefits nobody... and that is the reality for the vast majority of those hundreds of channels which are perpetually being multicast by cable companies. I get that the cable companies do not (choose to) perceive the additional cost from multicasting all of those channels... but there are indeed measurable costs to them, and one of those costs is the bandwidth usage of all those unwatched channels, which could otherwise be reallocated to help remedy that network clog issue that you mentioned. In addition, more and more consumers are already shifting their usage from the broadcast model to the streaming model on their own; it's not like the cable company is going to be able to stop the problem from getting worse. Thus, it is ultimately in their best interest to accept the inevitable, and find a way to reliably profit from the changing habits of their customer base. And -- aside from primarily sports enthusiasts -- their customers aren't really watching "live" television multicasts nearly as frequently as they once were. Thus, the case for multicast based systems becomes less compelling, every single day.
In addition to that, your observation brings up an ironic tie-in to the topic at hand, in that broadcast is an always-on technology which perpetually burns up power at both ends of the connection; thus, properly completing the transition to streaming will clearly save the cable companies some not-insignificant sum of money, in reductions to their own utility bills. (Not that customers will see that reflected in their bills, of course.)