I was a minor admin for Civcraft, and we went through something similar to, but less serious than, what you guys are going through now. After a year of runnning the Civcraft server with 300 players everyday in 2011 (or 2012, I forget), it was decided that Minecraft was too slow and too limiting.

We looked at Minetest, but it was way too rough at the time. And it still has security issues that will be difficult to address. The other Civcraft admins started developing a new engine. Despite having a lot of really good ideas, development ultimately failed. It turns out that programmers often don't make good businessmen.

I recommend setting up an email newsletter so you can inform players whenever you get Satlantis working on a new engine.

By line count, Minetest is written mostly in C++, so it will likely be a while before it could be made production ready. A complete re-write in an entirely different language may be necessary. I recommend writing such an engine entirely in a higher level language, such as Common Lisp. There are so many things that Common Lisp does for you that C++ and Java simply can't do without a huge amount of work (hence why Minetest will take a long time to be production ready, and why Minecraft in Java was limiting).

Then why not contribute to that codebase? Only by making it production ready can you then truly call it your game