In the past decade (i.e., when the money was plentiful) when a startup is young, the TC of its engineers rarely makes or breaks the startup. Being able to get an MVP out and iterate quickly is more important, so it was a rational choice to stay in the Bay Area even if it means 30% inflated TC. And after that moving is expensive in both time and money and risky (e.g., a key engineer might not want to go).
And having a critical mass of tech companies helps attract talent: if a company goes under or has large layoffs it is perceived to be easier to find a new job in the center of the tech hub.
I think covid helped nudge along the process of moving tech development out of SV, but it is a slow process. My 2c.
What has Meta or Google or Microsoft actually innovated on recently?
Google has a way of starting something and then shutting it down.
But there are frictions, too. Unless you go into management, comp tops out around $200k in most metros. HR -- instead of write-your-own-rules in a startup you have to take corporate training and get approval for things folks at the startups take for granted. Limited tools, externally managed corporate OS and software, Outlook instead of Slack. Office time requirements -- fully remote is very rare. And so on.
Not saying this is the wrong choice, just that there are tradeoffs.