Targeting cross platform anything drives the cost up immediately.
Just building websites that work on different devices if the site is non-trivial i.e. SVG, WebGL etc the cost up because each platform you have to duplicate the QA phase. This week I spent 4 days tracking down an iOS bug in a WebForm. The company had to pay me 4 days worth of work, half of that was reproducing the bug reliably.
People make out this stuff is easy. It isn't and I am sorry but Linux users on Steam is like 1 or 2% the last time I checked. It just isn't worth the effort for most developers.
Have you actually tried to do this yourself? With anything? I feel like everyone in this thread is trying to back-seat drive without realizing the effort that is actually involved in supporting multiple platforms.
We’re an embedded gaming system (casino) and I make sure we have a platform abstraction layer (2 c++ files per platform), and any project we build can be made for any platform (even Haiku). The amount of threading issues you discover makes this exercise absolutely worth it. The platform files haven’t been touched for almost a decade, write once (a week) and use forever.
We have and our software is better for it. Most of the stories about multiplatform being 'more expensive' or 'harder' are from corner-cutting devs and orgs.
It's nice to think that shipping a bunch of stacked hacks and making some money to then just get closed down/bought/retargeted is the 'happy flow' but unless you don't know any better you may want to diversify and find a job or project that has actual quality as a standard. I know that in late-stage capitalism with minmaxing in the mix that is hard to find in some countries, but it exists plenty if you go look for it.
Eliminate bugs, make releasing for PC and any console you want easy.
Linux shouldn't really be a special support case, it only becomes one when silly assumptions are made.