On rereading the thread, maybe I'm just misreading? I read your question as "they're popular, why do people think they are not" but you could also have meant "why doesn't anyone want them".
For the latter: because it's far higher friction than a phone call (or any similar tool). On the extreme end, I can walk into a meeting room and push a couple buttons and have a zoom meeting. And doing that with your computer or phone is significantly easier, often just one or two buttons whether you're a two person business or 200k.
Versus telepresence robots at the simplest: it requires charging, far more complicated UI to do anything that a video call cannot do, and is many times more expensive so you almost certainly do not have one everywhere you have a video-call-capable display. And the display is probably dramatically smaller, so you still need a separate display if you want to show anything useful. For very nearly everyone, that's just "a video call with extra baggage".
They can work just fine where those tradeoffs are offsetting vastly more expensive and higher friction things, e.g. in highly specialized surgery, and you do see them in those areas. That's just a rather small niche compared to "has a computer or phone".
Firstly: By the time your organisation is big enough that nobody will bat an eyelid at buying a $4000 gadget, it’s big enough you’ve got several buildings with several floors. Probably some doors too.
So you don’t need one robot, you need ten. And if it works really well and it’s a big hit? One per floor won’t be enough.
Secondly: The expense and maintenance burden fall on the recipient of calls, but most of the benefit is to the person making the call.
I benefit as the caller, as I can trundle over to someone's desk and interrupt them - but the benefit to them as the recipient is much more indirect.
Thirdly: Pre-pandemic, a lot of video call stuff was pretty unreliable, making the expense of a robot a risky matter. Post-pandemic, far fewer people are in a physical office - it's not like there are important in-office meetings that only have a single remote attendee.
I used to work at a research institute that had two campuses on either coast and we had those "iPad on a stick" type telepresence robots so people on one coast could attend (physical meetings) on the other. They eventually went away because more and more meetings either became virtual or were hybrid-virtual physical and so a laptop could work just as well as telepresence.
