Consumer/desktop audio on Linux will only advance (to the extent that it needs to - it already works quite well) because of people just sitting down and doing the work. Nothing is "driving" this forward.
The problems really arise at the junction between the two worlds: think bedroom/basement music/video production on your one laptop. This is the part that macOS/CoreAudio gets so right - there's no difference in any of this for consumer apps or "pro" apps. On Linux, you have to grapple with the awkward disconnect between these two "workflows". It's not THAT hard, but it's not THAT easy either. And this affects developers just as much as it affects users: which APIs to use?
Well, you probably need an application which has real-time requirements. This has always been the most difficult aspect of audio processing pipelines.
Real time audio programming within applications tends to be a problem, mostly because too many developers have never read or do not understand:
http://www.rossbencina.com/code/real-time-audio-programming-...
I haven’t exactly pestered everyone with a notepad and pencil taking a survey so I do not have a citation. I’d be VERY curious to hear a credible citation to the contrary showing that Macs are growing in use or even maintaining steady market share. That’s not what I see out in the world.
Ok, the situation in studios might be different (but I would curious to see actual numbers). In a live show you want reliable low latency audio (especially if you play VSTis on a keyboard or do any kind of live processing) and macOS is very good at this, better than Windows I would say. Dismissing this as "fashion" is a bit simplistic.
I'm a computer musician and audio programmer. Although I use Windows most of the time, I would be the first one to admit that the audio situation on Windows is a total mess. As Paul said, on macOS you have a single API and it just works (tm). On Windows, we needed an external company (Steinberg) to come up with a usable solution for low latency audio (ASIO).
Paul: I’m curious which “industry / workflow” is likely to be the best driver of improvements in Linux audio. Is film production or post-production the best bet?
(My assumption is that as long as musicians “grow up” on macOS and Logic and so on, you’re fighting with the “perennial year of the Linux desktop”. That wouldn’t apply to audio production teams who don’t need to also use their machines for their personal life)