Yes. But I suspect you don't understand precisely what that means:
First: A large number of mortgages were made, grouped into pools, and sliced into "tranches", where losses were assigned to the most junior tranches first, and the most senior tranches last. Thus, a junior tranche might stop paying out if 5% of the loans in the pool were defaulted on, while a very senior tranche might only stop paying out if 80% of the loans in the pool went bad. (Or whatever; numbers for example only.)
Second: These tranches were submitted to ratings agencies, who constructed models where they said "well, the chance of 5% of the loans defaulting is X%, the chance of 20% is Y%, the chance of 60% default is Z%...", then based on this model they assigned ratings to the tranches, reflecting the models estimation of their likelyhood of paying out, which is the entire job of a ratings agency.
Third: Tranches rated AAA were sold as AAA securities, because they were. The model said that Random Senior Tranche was 99.5% likely not to default, which is enough for a AAA rating from whomever was giving the ratings, so that's what it got. We've later come to question some of the models used, and in particular, it's been suggested that the ratings agencies incorrectly assumed default risk was less correlated than it was, and systematically undervalued the chance of a "black swan" event. But that's hindsight.
> If so, how is that not straight up fraud?
How is it even sort of fraud? Or if it is fraud, who do you think committed it, and what should they have done differently to avoid it?