Two very different things, but they've been (intentionally) conflated in a lot of the public debates to make it sound more evil.
As of now, all of that data staying safe in their solos means
(A) they don't lose customers who are fine with advertising but not that data going to some DB that anyone could then inspect (since you know some company would buy it then charge others $x thousands of dollars to dig into users' lives), and
(B) they're the only ones with that data, and thus their advertising business can use it to possibly show more relevant ads. If the competing ad company can just buy Google's user data, they could sell just as well-targeted ads as Google but undercut them by x% margin, thus always winning the ad space on a website.
Do you have hard evidence for this claim?
Where does one go to buy this Facebook or Google data? Who's buying it? Surely these things would have come up in all of the Congressional hearings, right?
[0] https://themarkup.org/privacy/2021/12/06/the-popular-family-..., https://themarkup.org/privacy/2021/09/30/theres-a-multibilli..., https://themarkup.org/privacy/2021/03/25/i-tried-to-use-the-... (more investigations have been done by WIRED as well).
Is that really any worse than sharing the sensitive data we store with them with known companies in ways that definitely compromise our privacy though?