That's sort of Google's reputation in machine learning: if you're tired of the lower salaries and grant-chasing of academia (or just don't like teaching), or have had enough of the uncertainty of a startup, Google is a comfortable place to go, with a large and stable salary, good access to data & compute time, good benefits, etc. Seems more common for people to try out academia and/or startups first and then go to Google later, rather than the other way around. I haven't done any kind of proper count or survey, though.