As far as BSD mindshare goes I think FreeBSD and OpenBSD have the largest pieces of a small pie. Meanwhile, despite how good their source code is, NetBSD still need to support so much esoteric, obsolete, and rare hardware.
None of this leaves enough mindshare for NetBSD to thrive in its niche.
This is, of course, not true. Most developers who support esoteric hardware do so because that's the hardware they like, or that they use, or both. Nobody forces anyone to work on anything they don't want to work on.
1.b) Yes, most new hardware aims to have Linux running on it as quickly as possible, but then what? There's lots of hardware that started with Linux support that never gets back in to the Linux kernel, so later either you're stuck trying to hunt down updated binary downloads, trying to find source trees that add support to the Linux kernel that may or may not work with newer versions of Linux, and/or running systems with known vulnerabilities that you can't easily patch yourself.
In many instances, hardware that has no forward path with Linux is well supported in NetBSD, because everything is in one tree and large changes that require specific hardware support updates aren't common.
2) You also seem to think that having code that supports esoteric, obsolete and rare hardware in the tree somehow introduces overhead. It really doesn't. This is a myth mostly repeated by people who've never really looked at the NetBSD's sources. Most old code can be left - people who want to run the hardware that code supports will let people know if something doesn't work. Again, nobody is forcing anyone to work on things they don't want to work on.
I don't think NetBSD needs to "compete" - it has users, both commercial and personal, and it has fans. No matter how many times people will announce that BSD is dead, so long as people are enjoying writing, updating and using NetBSD, it will continue to succeed.
