The most probable ‘worst case’ scenario is mainly: it turning into the same situation as mailservers. As the big brand instances enter the fediverse, the advertisement industry is going to go “hey, it’s all free real estate [for telemarketing]!”
We’ll end up with a fedi SORBS service that instances will periodically check reputation against for any ingressing federated activity. But we’re in a doomed time for anything like that to start anew, because I doubt most people can’t resist injecting their political beliefs in who’s considered a spammer or not, thus lacking neutrality.
In a semi-related example: even over in WordPress (but nothing design-specific to WordPress), people intentionally abuse the pingback functionality to promote their bullshit service/website/whatever, just as a means of “marketing!” Pingbacks were merely to be a social mechanism of “hey, someone mentioned your post on this website over here!”, but as with anything, just abused for marketing!
If there’s little-to-no cost to spam, then sociopaths are only going to spam to the extreme.
The only thing that would dissuade brands from touching it is: troves of completely uncontrolled illicit, deranged content that they don’t want to be in proximity of. YouTube content used to be a wild west, but then much heavier content moderation happened, then Content ID happened, then the brands and media outlets felt “comfortable enough” to move onto YouTube as their new home in place of cable media.
Alternatively, the only exit strategy is just burrowing deeper into the internet. I doubt telemarketers are going to follow people onto Tor and I2P.