The Linux ecosystem is by far the largest market for enthusiasts volunteering their time to develop a new browser. It makes sense this community would only care about building a browser for their own needs (Linux support only).
Additionally and unlike Mozilla, this volunteer community is also very unlikely to care about non-enthusiasts who may complain the browser doesn't support Encrypted Media Extensions (EME), Web Environment Integrity (WEI) or whatever anti-features ad-tech companies are trying to force onto mainstream users. This volunteer community is also unlikely to care too much about whether web sites containing 10MB of obfuscated JavaScript that was developed and tested solely against Chromium-based browsers works well. I think you'd find that the community would rather spend time working on projects such as yt-dlp to just re-implement front-ends for horribly broken websites, or would simply prefer to use non-broken alternative websites.
Linux is also the easiest kernel to develop against too for reasons that include _much_ better sandboxing features being available, better debugging tools and availability of source code to learn from and debug with. Contrast to Windows with undocumented or poorly documented kernel and other system library APIs, lack of source code (particularly examples of APIs being used in other software), and having to do more work to opt-in to security features that are enabled by default on a Linux system.
"smallest market" is relative, the Linux market for suckless tools for example is likely 10x bigger than the Windows one. For privacy focused alternative browsers I'd say its somewhere close to 50/50.
Which is not the point. If you want to have success you need to copy the Blender/Godot model. Year by year they make great versions for all platforms, and they do well. A new browser should support both and rally the people to work on it, again like the Blender Foundation has done.
Of course thats a great metric, but to better exemplify, taking Blender again, with their pool of money (mostly donated by enthusiasts) they could really accelerate the development, hiring the best contributors on a case by case basis, which solidified the code base and continuity. What I am saying is there is a great middle ground, with a good team, where money and enthusiasm go hand in hand.
Additionally and unlike Mozilla, this volunteer community is also very unlikely to care about non-enthusiasts who may complain the browser doesn't support Encrypted Media Extensions (EME), Web Environment Integrity (WEI) or whatever anti-features ad-tech companies are trying to force onto mainstream users. This volunteer community is also unlikely to care too much about whether web sites containing 10MB of obfuscated JavaScript that was developed and tested solely against Chromium-based browsers works well. I think you'd find that the community would rather spend time working on projects such as yt-dlp to just re-implement front-ends for horribly broken websites, or would simply prefer to use non-broken alternative websites.
Linux is also the easiest kernel to develop against too for reasons that include _much_ better sandboxing features being available, better debugging tools and availability of source code to learn from and debug with. Contrast to Windows with undocumented or poorly documented kernel and other system library APIs, lack of source code (particularly examples of APIs being used in other software), and having to do more work to opt-in to security features that are enabled by default on a Linux system.