Two competing theories here:
On the one hand, the single chokepoint people want a single defense point. They base their whole strategies around having a single chokepoint. With the current cap mechanics of a global cap on turrets, this effect is most efficent.
As a result of this, to play on a higher difficult then current skill, one can compensate by picking maps that aren't normal or realisitc. Tree, concentric, X, etc.
On the other hand, there are those who enjoy not having the 25 sq foot wall of death. They want more fluid games, where losing one world doesn't mean game over, and strategies are not on keeping that single world. They want a more "bend but not break" philosophy? They want a competing strategy. Competing strategies if somewhat balanced I think is good.
Follow the idea of the reproduction of the Z shredder. On a single world, you can build a cap of turrets with no problem. However, to build further then that cap, you face expodentally (x ^1.5) higher build costs and energy costs per cap. So to build twice the cap is three times as expensive, three times the cap is five more expensive and change times as expensive, etc.
What would this allow?
It would allow a great deal of flexibility. Having a single focal world would be more efficent then having 5 exposed worlds. Having 3 worlds would be more efficent then 6 open worlds. This would also make upgrading turrets viable. Getting those II's would make the base effect of the turrets 3x more efficent in your cap usage, and even more powerful then the current usage.