In this talk, we study the performance of regularized zero forcing (ZF) receivers as they apply to massive MIMO. Specifically, we consider two variants of ZF: regularized ZF and regularized ZF with box relaxation. For both cases, we derive precise error analysis at high dimensions using the Convex Gaussian Min-max theorem, a recently developed tool that emerged from the theory of compressive sensing. We take advantage of these accurate performance prediction results to optimally tune the parameters involved in each of these cases. In particular we propose an algorithm to tune the regularizer from a single observation when the SNR is unknown. Also, we derive the precise BER expression of box-relaxed ZF for BPSK transmission and use that to optimally tune the box parameter.