[Author here]
> To me, the number of Linux kernels running isn't the goal, but the benefits that Linux tends to provide in the areas of system openness
I think you're missing the points here.
The points are:
1. It's already won, on sheer numbers, or by unit sales, or by value of sales, or by number of users, or by almost any other metric you can suggest.
2. The war is not for the desktop and never was: that's a red herring. The desktop is one battle in a war, which is arguably for user eyeballs, or bums on seats. To win against an entrenched opponent with a secure base, *move the conflict somewhere else*. (There is probably some Sun Tzu aphorism about this.)
3. The real success story is FOSS over proprietary software. Software is just ideas how to do things, written down. When everyone knows how to do those things, the rivalry becomes religious not rational: "around here, we do things *this* way, not that way!" You can't readily defeat ideology with logic. But in the end you usually can with money: when it costs much more to follow one ideology than another, the expensive one loses out.
4. From #3 we can see that arguing intensely over ideological purity is fun but futile. It doesn't matter. What matters is now "how free is this?" or "how pure is this?" They are Jesuitical debates of little real significance. The real question is: do more people use FOSS than proprietary software? And the answer is now yes, they do.
5. The bigger point in terms of Linux adoption was in my final paragraph. That needs breaking down:
[a] Servers:
Those millions of Linux servers are mostly ephemeral. They are instantiated from a template, run, and then deleted. The distros and tools that make that [a] easiest [b] cheapest win, and nobody cares very much. Effectively, I suspect but cannot prove, most people run Debian in some form, but the less-technologically-adept companies run RHEL, which by the nature of FOSS pays for most of the R&D in Debian.
This is how FOSS ends up beating proprietary, and Linux ends up beating *BSD. R&D by one vendor mostly helps all the vendors.
[b] Client devices:
There are billions of Android phones in use. That means that *the majority of the human race* now communicates via devices running Android, and Android runs on Linux.
And the devices are lasting longer:
https://www.theregister.com/2024/01/23/second_hand_device_market/
What this means for Linux is that Linux *is good enough*. It works, it's reliable, most of the world runs on it.
But most of the world _doesn't_ run on Linux distros. They are not good enough. They are too hard.
The real question is: who will find a way to bring the reliability and simplicity of Android to desktop devices?
One day, somehow, someone will. And they will win.