I often tried to convince someone here, to help me with SNES classic mini, NES classic mini, Playstation Classic mini, and also the support for Raspberry Pi and the OUYA in this board is not as high, as I would like it to be.
There are about ~ 1000 people on Discord, who are loving the OUYA and also have an OUYA.
To be able to export to Windows Universal Platform, we can use App Game Kit Tier 2 C++. But C++ is also not as well documented and has not as much examples. Also the initial setup is not as simple with there templates we have to modify and so on.
So what we could do to make App Game Kit 2 more popular, we also would have to use more the C++ part of the API, I guess.
So let's say, there are about 4 mio. SNES / NES mini classic consoles, and maybe 1/2 mio. Playstation classic mini consoles out there and in use. Mostly are used hacked and for emulators.
And the problem with the OUYA is nearly the same, the offical support was shutdown. And for Raspberry Pi Zero / 3 / 4, there are often conflicts with newer OS-images or rights management and also some OpenGL issues because of the VideoCore4 inside the single board computer.
And all of these "consoles" have the problem, that non of them has a store like Nintendo, Sony, Microsoft or Valve have.
And also these consoles are more for people who don't want to use the internet, if they not have to. That is the charme of these retro consoles like the C64 and so on. But that also means, you would have to release your game physicly.
Also many of the Nintendo Switch games from indie game developers only have about 2500 cardridges, also if the downloads are much higher. And the next problem would be, to release a game, that is worth the release. Games with value. Not only a clone of pong, or some test or demo, or a prototype, but a real game, that has a playtime for some hours, has a unique visual style and so on.
And marketing would be a complete other thing.
But if someone of us would create a game in 2D with App Game Kit Tier 2 C++, I guess, it could be possible, to transform that into SDL, or some other libs and also convert it for SEGA Dreamcast or XBox classic.
I guess, we all could try the following:
OUYA (Android, console like, with a helpful community), hardware can be found for about ~ 25,- $ including controller, supported directly in AppGameKit with Tier 1 Basic, and with Android Studio or Visual Studio it is possible to use Tier 2 C++ and export that to the OUYA
a little bit more complex, but do-able, maybe with some more investment:
XBox One (S)
I guess, for about ~ 150,- $ or less. And this console could be switched to DevKit-mode and you can try your apps / games there, without the need for a paying fee like for iOS and Apple.
Also releasing on Itch.io would be possible for free, but is PC
maybe "SteamBox" with Linux, or what, with 100,- $ fee for Steam.
And it would be also possible, to get a DevKit from Nintendo for a little bit more than the price of the Nintendo Switch, but you have to have a game in the first place.
It is also free, to become a Nintendo Developer and for a fee you could buy a Nintendo 3DS DevKit or Nintendo WiiU DevKit. They are not as cheap as the other ways, but if you have your game in C++, it is compatible with there SDK and you could compile it for these consoles.
The thing with the MegaDrive / Genesis is, that you mostly would give your game away for free and or would sell it on Steam in combination with an Open Source emulator.
Also there are tools around for Home Brew, and the other thing is, that the consoles before the DreamCast mostly where very specific in there architecture.
Mostly games at this time also where converted by hand. Some routines had to be written again in another assembler and so on.
Also Sprites back than where very specific in size and color for each sonoles. And how many at the same could be displayed and so on. So also a conversion of a game, mostly was a new game, that maybe looked kind of familiar.
That is also the big benefit from the Nintendo Switch, PS4, XBox One, that the hardware is nearly a PC with a GPU from the PC. So you don't have to think about color palletes, you know, you have 24+8 (32), you also know, that your Screen can have 1920 x 1080, and that you don't have to could the lines and milliseconds for some screen flipping for timing. They have multiple cores and can do things in parallel. They can decode JPEG, PNG and MP4 and so on, they can playback MP3 and so on.
On the MegaDrive and the Super Nintendo the music was completly different. So you would have to compose new songs and so on.
I guess, if the games where made from a hand full of people in a few month, that could be done by hand. But know we have games, and also the costumer maybe would like to have those, that are made over years by more than a couple of people. So you could not convert the games anymore by hand or do this effort for every platform.
Also the question would be: Why?
Don't you have phones?
If you make a game, would you like to show it to the most people possible, or to a niche?
And does this niche only have this one channel?
The game collectors on YouTube like MetalJesus, JohnHancock and GameSack and so on, they have mostly all consoles. They would like to see your game on the MegaDrive, but they would be able to play it also on Steam.
Other people, who don't have all the consoles, and only have the MegaDrive and no internet, they also would never be able to play your specific MegaDrive game, because they don't know even, that it exists.
GameStop is or was shutting down, so even then, you would have to create a cardridge and make enough real hardware to sell it in these stores.
A lot of the OUYA games e.g. transitioned to Steam and or Android Playstore. Some made it to the nVidia Shield, some to AndroidTV.
What you could do, to make your game easyer for a convertion is, to have maybe 320 x 200, reduce your pixel art in the colors from the beginning and only use a handfull of sprites.
So it would look like a MegaDrive game in the first place, but it also could run later like that on real hardware. You have the assets, the sprites, maybe the sound effects and the gameplay logic.
This is a big amount on work you allready have done. So to convert that to real hardware, you maybe only need to re-write some of the 2D commands. Depending on how complex your game is, that could be done in some weeks, I guess.
You could create your game first for the PC and after you have something to show, you could search in the retro-boards, discord, facebook groups, and make maybe a Kickstarter and you will see, how much interest is there.
And if there is enough intereset, you could convert your game completly or pay someone who could do it.
Or you have to look into the homebrew scene and ask there for help and the right tools.