Favipiravir is a broad-spectrum inhibitor of viral RNA-dependent RNA polymerase (RdRp) currently being used to manage COVID-19 in several countries. By acting as a substrate for RdRp, favipiravir gets incorporated into the nascent viral RNA and prevents strand extension. A high mutation rate of SARS-CoV-2 RdRp may facilitate antigenic drift as an answer to the host immune response, thereby generating resistance of virus to favipiravir. Therefore, it is extremely crucial to predict potential mutational sites in the RdRp and the emergence of structural modifications contributing to drug resistance. Here, we used high-throughput interface-based protein design to generate >100,000 designs and identify mutation hotspot residues in the favipiravir-binding site of RdRp. Several mutants had lower binding affinities to favipiravir, out of which hotspot residues with a high propensity to undergo positive selection were identified. The results showed that the designs retained an average of 97 to 98% sequence identity, suggesting that SARS-CoV-2 can develop favipiravir resistance with just a few mutations. Notably, we observed that out of 134 mutations predicted designs, 63 specific mutations were already present in the CoV-GLUE database, thus attaining ~47% correlation match with the clinical sequencing data. The findings improve our understanding of the potential signatures of adaptation in SARS-CoV-2 against favipiravir and management of COVID-19. Furthermore, they can help develop exhaustive strategies for robust antiviral design and discovery.