Consider this: days after the successful sequencing of Omicron, similar cases were reported in dozens of countries, having arrived there just before or just after the announcement. We know, that it's "just before," because references are being kept and re-analyzed in those cases, and none of the countries with the resources for genomic surveillance has any data on earlier strains in the phylogenic tree.
It is simply very unlikely (not impossible, but if you're in London and hear hoof beats, it's probably a horse, not a zebra), that Omicron could have developed over 33 N and a superset of S mutations without ever popping up in any surveillance. South Africa, especially, is a key surveillance player. They're often much more granular than even European key players like the UK, Germany, or France.
A strain with this level of infectivity doesn't linger in one hospital. From the looks of the genome, that part (infectivity) would have developed before immune evasion. Having a highly infective strain evolve would ring alarm bells in labs the world around.
I guess it's pretty non debatable that Omicron started its World Dominance Tour in October or November 2021. To get there, it had to undergo a massive sequence of events making it as far removed from α.7 as it is. So we can state, with some certainty, that it evolved in isolation and broke free shortly before being discovered.
If we now consider the possible origins, wild zoonotic, single origin immune compromised, or accidental leak from research, the latter doesn't sound so improbable anymore. Again, horse, not zebra.
