And we face situations like this all the time as developers building full-code solutions. We get stuck because the PoC which went into production can't scale, and we buckle down and build an appropriate V2.
You're essentially arguing that MVPs are a bad idea, but evidence suggests the contrary.