The long answer is that if you have mods, you'd have to develop those mods, and any player that wants to play it will have to get said mods. It'd be a massive drop in players, because not everyone that wants to play will be able to run mods with custom models. You would also have to update the server and the mod, while developing both, spending more time on both, and over time that will become harder and more tedious.
The fix, or alternative, would be like what Wynncraft does, a texture pack with custom items/item models and plugins. This is, however, very much so limited, and then you'd have to (if you didn't make the plugins) wait for those plugins to update when the game updates, as well as the texture pack if you want to add more items, etc. It would, however, mean that all people need is vanilla Minecraft to play.
Shortened down, it'd be a pain to keep up, and that can make them lose players more so than is seen as necessary or reasonable. This can also mean that anyone that relies on it as a job (as some people do if developing/working on a server takes up too much time) is at risk of losing a lot of money, and that can mean losing their lifestyle.

So, theres that, or developing a new game they can control, add what they want, etc, without having to run the strain of keeping up with updates on a game you can't control much of other than what version they're on and how updated their own mods/plugins/etc are. So, it isn't necessarily fixing something that isn't broken, its more like an in depth customization to something they may be able to better work with, if that makes sense. This also means any developers using it as a job will not lose as much because "I want to play 1.15/1.14/w.e update, but [Server] isnt updated. Oh well /moves to different server, possibly never comes back"