I will say that US 5.56 is on the edge of pressure limitations based on temp variables, and is at max for a mass-produced standpoint. Since the US ammo is oriented towards the DI gas system of the AR15 family, it wouldn't surprise me that the gas pressure curve wouldn't run the L85 well.
The guys in SAS and SBS use the Colt Commandos because it's the best assault rifle platform in the world right now. If anything better would come along, they would have it.
With assault rifles, the current trend is seeing higher capacity magazines being developed, such as the AK74 quad-column mag, the SF60, US Palm, and the as of yet unseen Magpul 60rd PMAG. I think it would be a mistake to reduce the rifleman's mag capacity from anything less than 30rds. I also feel that it IS possible to make a 30rd Grendel mag that is the same size as a 5.56 30rd mag, but I still think we need to be cautious about using the Afghanistan AOR as a major input for the next assault rifle design as a primary consideration.
Also, regarding the yaw performance percentages, what is the expectation? For a projectile to behave predictably 100% of the time in a mass-produced assault rifle load? We have to establish some expectation guidelines before citing percentages for how many times a certain projectile does X, Y, & Z in ballistic gelatin. 70% consistency seems pretty dang good to me, for a standard AR load. From a soldeir's perspective, I'm not waiting to see what just one round does to an enemy, regardless of what rifle or carbine I'm carrying. They will get at least one follow-up shot at distance, and 3-5 at CQ ranges, unless a head shot drops them like a sack of potatoes before I can make another at distance.