Archeage an MMO built using cryengine.
http://s6r5v2u5.ssl.hwcdn.net//120/NHP/CUL/172/U9C.jpg
https://en.wikipedia.org/wiki/ArcheAge
The Cobra engine was not designed for MMO's either. The point is that it is possible to do it and other companies have used cryengine to make an MMO. This means that it was possible for CIG to use cryengine to make an MMO, they just do not have the talent to do so.
CryEngine was not built for MMOs.
It can be adapted for MMOs...as ArcheAge suggests. I think I even posted a quote from the Aion development team telling us how much of CryEngine they had to rewrite to make it suitable
Whether he chose Unity or Unreal or CryEngine....Chris Roberts was going to have to do some serious rework of the engine to make the MMO style game he wanted work.
The problem as that he was stuck with CryEngine, and while that was perfectly satisfactory for the game he planned, it needed major reworking for the game he was actually making. Unlike other teams, such as those who developed ArcheAge, he didn't do it.
Or rather...he did do it, but he complicated matters (putting it mildly) by developing the game at the same time. That creates all sorts of nasty issues.
The ideal solution would have been for Chris to go to CryEngine and seek a termination. He was now developing an entirely different game, which required new testing and feasibility studies, new planning and so on. CryEngine was no longer viable...no engine was...and he needed a custom engine.
The next best solution...probably what he was stuck with given the GLA....would have been to recode CryEngine into the engine he needed it to be, and to do so without the complications and problems that would arise from simultaneous game development. The downside is that there would be little or no extra fundraising. The upside is that actual development, once started, would be smoother, quicker, easier, cheaper. Better AI, 64 bit positioning, localised physics grids...there was nothing new or impossible; it just needed to be added to the engine. It didn't even need to be "perfect". Once the basic structure of these new features is added, they can always be refined and expanded and optimised later, when the game was under development. But the game developers would need to know what the engine needed and what it was capable of. There's no point developing a 64 bit positioning system if the engine will only accept 32 bit co-ords, and there's no point developing the system until the engine can support that. All that would do is require you to recode the game as well.
The worst solution would be to develop both the game and the engine at the same time and accept the inevitable delays, problems, complications, bugs and performance issues that were certain to arise by trying to build a foundation after the house had been built. It's not impossible to do - but it's easy to see why everyone else does it another way. The only advantage this option offered was that it allowed CIG to continue fundraising by allowing them to get something visible out the door to show backers quickly.
I've said it before...any engine can be used to create any game. It really depends on how much money and time you are willing to invest to account for the shortfalls in engine capability.
CE could easily be adapted to support an MMO. It's been done. New netcode and all.
CE could be adapted to have localised physics grids and subsumption AI and all the rest. None of this is new, none of this is ground-breaking - the effort would be adding it to the engine rather than coming up with anything "new".
But it is a LOT easier to modify the engine and get it working if you don't have to worry about the developing a game at the same time. And this isn't something you can just palm off on another team. The engine SHOULD be ready - at least, the skeleton of the engine should be ready - before actual game development takes place. You can flesh out and improve the engine, but you can't make wholesale fundamental changes to it without impacting on game development and vice versa.