But for cellphones, lots of values of "niche audience" end up with, "pick zero." Here are some examples:
1. Hook up a keyboard to ssh into the device. 2. Hardware switch to turn off the baseband OS. 3. Software toggle to turn off the baseband OS. 4. Sandboxed baseband OS that isn't proof-of-concept or impossible to buy and use in the U.S. 5. Free software baseband OS. 6. Open source "tivo-ized" baseband OS. I.e., user can read the source but cannot run modified version. 7. Just glue together a damned RPI with a baseband OS and ship it to me in working order.
Those are just the niches that interest me. Only #8 looks feasible in the near future and its apparently still in the planning stage:
https://www.crowdsupply.com/arsenijs/zerophone
So I think your "niche audience" needs an asterisk that says, "most niches not yet supported."
Also, you're missing #8.
I am not active in the porting side right now, but am prototyping a mobile computing device that runs pmOS. It's kind of a neuromancer style ono-sendai portable deck: A rectangular box (approx 11x29x3cm) that you can opem up and velcro your linux phone into and have a full-sized, stainless steel, porclean, or plastic mechanical keyboard and foldable mouse with slots for extra memory and battery life. It is designed to be as compact and as durable as possible, while staying true to the postmarket name by sourcing from reused materials whenever possible.
My hypothetical market is mainly highly mobile autonomous individuals residing in developing world megacities, but I'm interested to know what someone from the US or Europe thinks about the idea.
Oh ya, one more thing, the profits go to pmOS, to scale up a re-wilding project, and to getting these hackable linux devices into the hands of children forced to work in the supply/waste chain of electronics manufacturing.
How's that sound?