See, I thought it was probably a good deal for Amazon because I think that many people who do get the subscription will not get more than a book a month, in the future. They will when the subscription is shiny and new, but eventually they'll go to whatever their book reading habits were, and I'm sure Amazon's tracked that enough to know whether they'd make money on this. For example, I read 3-4 books a month on average, but I don't read them all on Kindle because I will always be seduced by a beautiful cover or by a convenience buy in person. Furthermore, a lot of books I read are rereads, particularly as a later book in a series comes out. There are times when I go through a Kindle-buying phase, and times when I go through a library phase, and times when I'd rather have a hardcopy in my hands. Most people don't read that many books a month, but just like many people will pay for a gym membership because for one month they'll go every day and the other 11 months of the year they'll pay the gym for the hope that they'll go again, my guess will be that some people will pay for it long after they establish that they don't use it enough to justify it.
But all of that also assumes that the books that I would want to buy on Kindle are the ones that are available for Unlimited. If you're stuck with self-published books, they're probably going to make a killing. SP'd books tend to be cheaper in the first place, so if you spend $9.99 and get 2 SP books that were $3.99 each, you're spending more that you would have. If you don't find many SP'd books you're willing to read (even for "free"), your subscription is paying for nothing. If they're paying their SP'd authors less for KU than for purchases, they make even more.
And I'm sure they thought that they'd get more big publishers on board when they started this enterprise. They've been in talks with some of the publishers lately, and I wonder if this was a major subject.