Can we save money by migrating Oracle to Postgres?
Probably not, so you should consider easier ways to save money on Oracle.
Saving money on Oracle software rarely has anything to do with using less of its database. Start here to understand why: not only repricing, but also matching service levels and license sets, policies that lock customers into paying annual support for decades.
Postgres is hot though, similar to MySQL many years ago before Oracle acquired it. AWS and a myriad of systems integrators advertise savings opportunities associated with refactoring Oracle workloads to Postgres and other open source solutions.
Opting for Postgres over Oracle for a net new implementation should avoid cost because Postgres is free. Whether implementing and managing Postgres is less expensive is another matter, unaddressed here.
Moving an existing, licensed and supported Oracle workload to Postgres to reduce annual support is a heavy lift. It’s far more likely that commoditizing Oracle to low core, high clock rate chips, running single instances of 12.2, maintained from an archive of patches, with no support from Oracle, is the most effective solution to saving money.