How is that tautological? It's clearly an empirical statement. My source is my own observations of decacorns as well as IPOs during the last year.
That said, feel free to skim through this (slightly outdated) list and note how a large fraction of the US-based companies are based in SF or SV: http://fortune.com/unicorns/
Skimming through that list they look like long-running businesses, not startups. They may have been startups once upon a time (same as any publicly-traded business for that matter), but they're completely irrelevant to a discussion of where modern day startups seeking to scale to billion dollar entities should be founded.
That said, feel free to skim through this (slightly outdated) list and note how a large fraction of the US-based companies are based in SF or SV: http://fortune.com/unicorns/