Maybe I am being too simplistic here. But I have never received a spam message to my XMPP account and I don't know how a spammer would find it.

In a phone-based system a spammer can spam a list of numbers, or use contact lists that are easily shared via phone permissions. There are several low-effort discovery processes.

For e-mail, you get spam when you you input your personal e-mail into forms, websites, or post it publicly.

But for something like XMPP... It seems rather difficult to discover accounts effectively to spam them. And, if it is an actual problem, why not implement some kind of 'identity swap' that automatically transmits a new identity to approved contacts? A chat username does not need to be as static as an e-mail or a phone number for most people.

I just don't see 'spam' as such a difficult challenge in this context, and not enough in my view to balance out requesting a phone number. Perhaps a spammer can chip-in?