I think it's just simpler than that, it's risk. Some content has a known level of "bad" risk associated with it, legal pornography for instance, but other content potentially has a massive and unquantifiable downside almost to the point of being an existential threat to a platforms business.

You can't force people to host content or be associated with hosting content which they fear will blow up in their faces.

So far almost all private censorship I've seen can be chalked up to profit motive. Payment processors don't like adult content not because they're prudes, but because there's a significantly increased risk of chargebacks, so why bother. Enabling illegal activities while maintaining plausible deniability and making easy profit off it? Why companies wouldn't do that? On the other hand, being a neutral platform is only fun until the Twitter mob notices you, after which you can choose between losing neutrality or getting painted as a bad actor on the front page of newspapers (sometimes you get the latter anyway, which incentivizes others to abandon neutrality preemptively). And even if you don't care, advertisers do care, because getting a mass-market product ad next to politically incorrect content may cause them issues with the Twitter mob and front page of the news.