To be sure of some of these things, we'd not only need more data but to be able to see the future.
We've had some pretty awful times in the past. The era of the robber barons comes to mind. At that point, people didn't have a concept like late-stage capitalism. But if they had, surely they would have thought they were living in the late stage. They would have been wrong. In the immediate aftermath of the crash in 1929, surely some people would have thought they were living in the aftermath of the collapse of capitalism. They too would have been wrong.
We could be headed straight into catastrophe--or not.
Both of these historic periods led to some fine-tuning of capitalism and ruled out certain kinds of disasters. For instance, since the SEC opened its doors, we haven't had another crash comparable to 1929. Some people would see these adjustments as half measures at best, and human greed will find a way to exploit whatever system is in place. Still, we are arguably less vulnerable on a number of fronts than we were in 1870.
The Cuyahoga River? Yes, something should have been done about that earlier. We forget easily. It's like people who keep rebuilding in areas prone to wildfires during dry seasons. (Though in that case, one could argue that they're stuck. Who is going to buy their property in a wildfire zone? Can they afford to just walk away and buy somewhere else?
KDP is a different situation. If you're losing money daily, you kind of know it and are constantly reminded if you don't. If your books aren't selling, you definitely know it. Unless you're visualizing a huge number of people who are independently wealthy (and they're aren't that many), you'd have to admit that authors can't maintain themselves in those circumstances for very long. You might leave your titles up, but you aren't going to keep paying for AMS ads with money you don't have. Amazon may at first have boosted income from AMS with its various shenanigans, but that just isn't sustainable.
As far as how much Amazon will lose, it's true that books at 28 billion (if the stat is from the same year as yours) is only about 5.5% of the total. If Data Guy's old figures are still roughly accurate, indies would be somewhere between a third and half of that. Let's say it's 10 billion. Know any companies that will basically throw 10 billion away for nothing? No, neither do I, and it cuts against your greed analysis of motivation.
Let's not forget KU, which in the breakdowns I've seen, is listed under subscriptions, not sales. Because a lot of trads still look down on KU, indies are almost certainly at least half of the KU income, maybe more.
If Amazon really didn't care about the income it gets from indies, who maintain the exclusivity requirement on KU? It costs some resources just to monitor and make sure our KU titles aren't showing up on other sites. It also costs Amazon a lot of titles that could otherwise be in KU--again, less money for Amazon. In the beginning, Amazon was trying to pull us away from the competition. But now, there isn't all that much competition in books. The mom-and-pop stores are all gone. The big outlets have only a small fraction of what Amazon has. (When I started, Barnes and Noble had 20% of the US market. Last time I checked, it had 3-4%.) The motivation often alleged for the exclusivity requirement no long applies. Yet Amazon keeps it in place. The only logical possibility here is that Amazon values the indie herd, at least a little. Their current policies, however, do not well support such a valuation, leading as they will inevitably to declining sales, declining KU revenues, and even declining AMS revenues.