Or, what is more likely, we will find a more recent ancestor strain to SARS-CoV-2 that has 99%+ identity to certain ORFs/contigs/parts of the genome. (For spike, we already did with the Banal sequences)

Invoking a reverse-engineered synthetic patchwork assembly of already highly similar sequences (some of which were sequenced AFTER SARS-CoV-2 emerged, but hey, details) just seems like pressing a square peg through a round hole.