Your point reminded me immediately of this article from late last year:
https://redmonk.com/sogrady/2018/12/21/cycles-oss/
The article's reasoning struck me as fundamentally sound, but it began with the false identification of open with the public, rather than private, interest.
As for your proposal, have a look at the Fair Source License: https://fair.io/
I blogged initial thoughts on Fair, line-by-line, when it came out: https://writing.kemitchell.com/2016/03/30/First-Read-of-the-...
I think Stephen probably coined the "There is no open source business model" expression a couple of years back.
I think what really happened is people didn't quite realize, or didn't think it would matter, what they were giving up when they open sourced their software. It sounds good and feels good but 5-10 years later when it's a core component at major companies it feels like you got ripped off even when you explicitly signed up for it.
You could frame this in a different way and come to a different conclusion. Instead of describing OSS as a "stack you lose the license for if you do too well" you could say "stack you don't have to pay for unless you're successful".
A similar question exists for laborers that need to invest in learning tech stacks. Why spend time mastering something that certain companies are discouraged from using? Especially if a non-encumbered variant exists.
Wouldn't it be more practical to learn tech that employers can easily use?
I'm not convinced personally, but I think that's the idea.
However, I understand the point being made. It's essentially a tragedy of the commons argument. Because the companies that get the most value from open source are also often the ones with the greatest resources to give back to open source projects if they chose to, they have a moral obligation to do so. There's also a sentiment one often hears along the lines of: I'm fine with people using my work for free but I don't want them making money off it. (This is essentially the basis for Creative Commons' non-commercial variant).
So it's not so much that large companies are actively hurting the projects--though you could argue that hosted offerings compete with the projects themselves. But that they often don't help in anything like the proportion that they benefit.
So all the examples you give are funded by big companies. Works perfectly.
In the end, they probably should have had those provisions to begin with, but were more concerned about early growth and first to market.
That's only sustainable for as long as there are those comparable and freer platforms to switch to. But if the argument being made here is valid, that is not guaranteed.