That is a problem, if they do optimization before new features they will be forded to do rework.I don't think I would be happy at my work if my boss said "Go optimize the data handling over a network interface for project X, BTW, we are updating features all the time so its a moving target. Good luck!" I think I would give him the crazy eye and try to wait 10 seconds before I point out the tremendous waste of time he inflicted on me and the waste of resources that was about to happen.
That is a very specific science to optimizing network use for an application. Part of that science is the application can't keep making large changes beyond bug fixes. You would spend a week getting a baseline, planning changes, identifying bottlenecks, and then toss out the work to just start over the next week when they add in a new big feature.
TL;DR - Optimization comes close to last in a large project like a game. Be much more patient.
If they do new features with out optimization it will be hard to know if the features works the way it should be.
That is the case, today the servers cant handle even an small war between 40 players, its possible to optimizate it to allow this kind of game ? If yes, ok, if no, then new features that have battles like in mind that will fail because of performance...
If they have no awnser, they have the risk to build features that will never work, so its lost work anyway.