I think the assumption would be more that FAANG would throw wads of cash at non-FAANG candidates than the reverse. There's a few "common knowledge" things (assumptions, but often accurate) about recruiting from FAANG that would certainly explain the lack of "wads of cash":
* Golden handcuffs are real and deliberate and in many ways rational: very few other companies have similar-sized wads of cash to throw OR similarly-scaled problems where FAANG-level salaries are likely to produce significant ROI
* FAANG stacks tend to have a lot of amazing in-house niceties that can't be found elsewhere, so there's a certain "are they an expert or are they just an expert at using Google's internal stack"? Would they be productive out of the gate or would they need to rebuild all that stuff first?
Everybody wants the prestige of telling their CEO or their investors that they poached people from FAANG but few have the wallet or the precise scenarios to get into a bidding war about it.
I had these issues as a HM in 2020 trying to hire ML engineers. (At mid-size places there's also often a "WTF do we do with a ML engineer" problem where if you aren't a data scientist or a general SW engineer, it's hard to carve out a role in between. Especially if the org has data scientists who aren't particular "engineer savvy" in terms of workflow and tooling. And then the ML engineer makes things shiny and version-controlled and all... and the data team mostly just ignores it since moving over to that stuff isn't one of their KPIs...)
* Golden handcuffs are real and deliberate and in many ways rational: very few other companies have similar-sized wads of cash to throw OR similarly-scaled problems where FAANG-level salaries are likely to produce significant ROI
* FAANG stacks tend to have a lot of amazing in-house niceties that can't be found elsewhere, so there's a certain "are they an expert or are they just an expert at using Google's internal stack"? Would they be productive out of the gate or would they need to rebuild all that stuff first?
Everybody wants the prestige of telling their CEO or their investors that they poached people from FAANG but few have the wallet or the precise scenarios to get into a bidding war about it.
I had these issues as a HM in 2020 trying to hire ML engineers. (At mid-size places there's also often a "WTF do we do with a ML engineer" problem where if you aren't a data scientist or a general SW engineer, it's hard to carve out a role in between. Especially if the org has data scientists who aren't particular "engineer savvy" in terms of workflow and tooling. And then the ML engineer makes things shiny and version-controlled and all... and the data team mostly just ignores it since moving over to that stuff isn't one of their KPIs...)