I've been running what I believe is the same Norma brass from MidSouth that you have. According to them, it is made for them by Norma. I also have fireformed all the pieces once with a reduced load, and have just begun to work up full-pressure loads with it, so mine is only 2X-fired at most. As yet, I have not observed any issues with it. However, I am using it in .223 Rem F-TR rifles, which are not AI chambers.
As XTR noted above, case-head separation is typically a totally different issue than simple over-pressure, which usually kills the primer pockets in 3-4 firings with the hot loads we use in F-TR (2800+ fps with 90 VLDs). As I understand it, case-head separation in a .223 is far more likely to result from over-working the brass (i.e. too much shoulder bump over 2-3 re-sizings will often cause case-head separation). Is it possible that could have occurred somehow during the AI fire-forming process?
Another possibility that occurs to me is that on the surface, a load with 24.0 gr Varget under 88s seems like it ought to be generating a lot greater velocity than 2640 fps/44.7K psi. However, I don't know your barrel length or freebore length, so that is merely an observation. Nonetheless, it's probably worth re-checking your QL file to make certain all the inputs are correct, and the QL pressure prediction is really matching up like you think it is. It won't cost much effort just to double-check and be sure.
Finally, you might also want to compare you shoulder measurements (i.e. how much the CBTS measurement grew after each firing). If the growth prior to re-size is excessive, it could be a over-sized chambering issue, which could certainly be associated with case-head separation under the right circumstances. Again, I can't tell how likely this would be, but it's something you can easily check with little effort on your part, which may at least allow you to rule out one possible cause. Sometimes when you experience issues such as this, examining each possible variable one at a time is the only way you can ultimately find the cause. It can sometimes be painful, but solving the issue makes it worth the effort.