The theoretical analysis of R-fold modular redundancy, cascaded R-fold modular redundancy and NAND multiplexing is presented and these fault-tolerant techniques are compared in terms of resistance to massive levels of defect density. Optimal cluster size analysis and redundancy optimization of the cascaded R-fold modular redundancy technique has been performed for the first time in the context of a large-scale system. The optimal window of application of each fault-tolerant technique with respect to defect density is presented as a way to find the optimum design trade-off between the reliability and power/area. Building viable systems consisting of components with high defect densities in future nanoscale technologies will have a high cost in power/area, regardless of the fault-tolerant techniques used.