Your point is completely valid and it raises good concerns. In this case, I think it’s worth separating AMP and Google’s implementation.
Google’s dominance means that content distributors will build to AMP spec, and there’s suddenly a world where “tag your shit” is taken seriously, and that content/those feeds exist in the open where anyone can download them.
If you’re a non-Google tech giant you’re going to drag your feet on AMP. You can make publishers dance to your tune.
But for the little tiny mammals that we are (I.e. Postlight) I just see a huge outpouring of valid, parseable, usable HTML-ish stuff that is well-tagged showing up in the wild. I don’t really care about the motivations of the publishers or Google. I just know that there’s suddenly, apparently, 150 million data-rich URLs with valid markup on the open web that could—if I had a good parser—be reliably parsed by machines and work on mobile. Independent of Google.
Plus it’s just useful to have a format that works that well.