Drug resistance in human immunodeficiency virus (HIV) continues to be a pervasive problem affecting the lives of millions of patients worldwide. In this study, using a Potts protein sequence-covariation statistical energy model combined with Kinetic Monte Carlo (KMC) simulations of evolutionary trajectories under drug selection pressure, we explore the evolution of drug resistance in HIV, starting from an ensemble of drug-naive patient protein sequences. We follow the time course of a total of more than fifty drug resistance mutations in Protease, Reverse Transcriptase, and Integrase. The KMC model on the fitness landscape accurately captures the relative acquisition times of drug-resistance mutations reported in the literature. We find primary drug-resistance mutations with long acquisition times typically require a larger number of accessory mutations to prime the sequence background for their acquisition and eventual entrenchment, while those with shorter acquisition times exhibit pre-existing high initial biases within the drug-naive patient population. We introduce a new metric, the potentiation factor in drug-naive patient protein sequences, that can be an accurate predictor of acquisition times for mutations subject to external selection pressures.