This was discussed, a few years ago. Nothing came of it (as expected).
IMO, trying to build retro hardware is certainly possible. The cost would be far more than anybody would ever consider paying. None of the required hardware is currently in production, so it would mean paying manufacturers to retool to produce ancient hardware. They would do it, if you pay them enough, but I guarantee they won't do it for small lots. The benefit of making retro hardware is far from certain too.
Better to put the money into making current hardware work. Arca Noae (and partners) is the only company in a position to be able to do development, but they are under funded, and under staffed, so development is far slower than what most people want. The important stuff does happen, eventually.
I don't think there will be any real improvement in the software until there is control of the source code. Everything else is a hack.
IBM has made it very clear, that there will be no source code (for many reasons). They too would release it, if somebody wants to pay them to do it (probably a billion dollar, 3 year project, with very questionable results).
OS/2 is what it is. IMO, it is still more capable than win 10, if win 10 had the same restrictions (32 bit, lack of drivers, etc.). The biggest "problem" is that very few, if any, developers are doing native OS/2 programming, and software from other sources has become so bloated that it doesn't work very well (even on the native platform).
Support the developers, and we will have OS/2 around much longer than many people seem to believe. One advantage is that you can buy surplus, or used, hardware off the bargain shelves, and it is likely that OS/2 will work. Of course, you do need to be careful what you buy, and paying good money for some exotic video adapter is not smart.