Looks like they have genomic enhanced epds from dna. Is that correct? I don't know how AAA uses dna in regard to epd's. But the simmental association calculates epds at the time of registration using the parent's numbers and whatever data was submitted on the calf (bw, ww, etc). Then dna can be submitted for enhanced epds. When the dna data is received, epd's are updated to include the genomics. You can then compare the initial epd's to the new enhanced ones to see the amount of change. Sometimes it is considerable.
I think the result of genomic enhanced epd's is better accuracy, not necessarily better epds. Some epd's may go up and some may go down. If the epd's don't change after the dna work, then there would be no purpose in doing it. Look at the changes and differences as better information for selection, not as a mystery. I don't think it is unusual to see considerable variation in progeny of flushmates - with or without the genomic epds.