A quick search claims FB's annualized average ad revenue per user for 2018 was just under $25. If I got value out of FB (I used to be a regular user, but now I just have the account for event invitations and otherwise never check it), I would absolutely pay $25/year to be able to use it without ads.
I suspect FB is near the top end for how much ad revenue they're able to extract from people; I imagine most sites are sub-$1 (certainly sub-$5), and could price subscriptions accordingly.
Only a small subset of users can/would afford to pay $25/yr for Facebook, and these users are arguably the most valuable to advertisers. Therefore by letting them skip ads, FB would lose out on more than $25/yr/user.
It might be profitable at something like $1,000/yr, but who would pay that?
And besides, there's always a number that works out to a break-even point. Maybe it's above $25/yr, but suggesting it's as high as $1000/yr is absurd. Even if it's $100/yr, I'd still bite (not for FB, but for a site I'd actually get use out of to the degree that many people do with FB). Sure, that further reduces the pool of people who could/would pay, but it'd still be an available option for some people.
So you're paying to remove ads but tracking and monetisation is still happening.