Complex system implementations combined with the latest technology nodes allow us to implement hardware for versatile applications. The ever increasing demand for quick time-to-market has led to the widespread use of Intellectual Property (IP) in ASIC design methodologies. These developments, in addition to manufacturing limitations, make early prediction of manufacturability for complete systems challenging. We present MIDAS: a scalable, IP-inclusive model to predict system manufacturability. Results from applying MIDAS to an embedded processor system reveals that several useful insights can be gained towards realizing yield budgets for complex systems allowing quicker co-optimization of all implementation goals.