Stuff for your personal cellular use would never come close to covering the costs involved. And 4g will still be used for many years to come for that.
5G covers 3 variants: 1) massive broadband (with mmWave in particular); 2) URLLC - Ultra Reliable and Low Latency Communications; 3) massive IoT. The current focus is on (1). Lot's of talk on (2), but nothing much concrete yet. (3) is moving along, but it's actually based on... LTE.
Strictly speaking, 5G is a set of requirements defined by the ITU-T, not a technology. The actual technologies are developed by another organization, 3GPP. And there are 2 technologies to cover 5G: 1) NR, or "New Radio". This is what most people mean by 5G. It's for massive broadband and URLLC; 2) LTE (release 15 and later) for massive IoT!
Yes, the IoT variant of LTE (LTE-M and NB-IoT) will be the 5G implementations for a while. Eventually there will be new NR versions for IoT, but nobody is in a hurry there. LTE-M and NB-IoT evolutions will be just fine for a long time, as far as massive IoT is concerned.
When you hear about 5G deployment today, it really mean "NR" and not LTE based IoT. The concern for smartphone is really capacity during peak time in the busiest cells.
I work in the field BTW, as you may have guessed ;)
I personally think the entire mmWave part of 5G is at least another 10 years away. Like, that's just going nowhere. Massive bandwidth? Great, what for? WiFi will be mmWave before we have an application for that in cellular.
Presumably what is actually being deployed is the incremental LTE revision on the same old bands, some extra thrown in that were yoinked from other applications.
My main exposure of the 5G area was the parts around the move to Network Function Virtualization. I wasn't deep in enough to know the ins and outs as you clearly do, I was extrapolating from what I know.
Here's my reasoning, feel free to poke holes.
Background: The hardware boxes used for network functions supplied by Nokia, siemens and the like were very expensive.
Telco profit margins have dropped massively as people make less regular calls and SMS, and use whatsapp and viber (over the top) instead.
The market is saturated.
A) Making data "Better" makes these over the top services better. If my data is more stable I will more likely make a call using whatsapp, and the telco looses the money from voice. - why would they do this?
B) 5G requires completely new infrastructure, from aerials to core packet switched. This costs a lot of money, again, why would they do this?
Reasoning: The ITU defined the whole NFV thing, and how to do it. You need a certified platform, cloud orchestrator, a whole load of systems to do NFV. The move to NFV is very similar to how x86 went from bare metal in the data center to hypervisors. It is an absolutely massive change. The telcos have never virtualised their machines, and can only start doing it now (in the last 2 years)
This is why the telcos took on/over openstack as a concern. I think they thought the could run openstack clouds and run their NFVs (this is what telcos call their VMs) for way less money than the hardware boxes.
Of course, reality came home, and rolling your own is hard. And the NFV providers still charged a fortune, and so did the system integrators.
So again, we are at a point that this costs a load of money for the telco. Why do it?
Benefit one is that they can now roll out new functions in a relatively short amount of time (magic of virtualisation and all that). And by short I mean I have seen a national VoLTE roll out in a few weeks.
And I though, that's great, but me, as a consumer, am I going to pay extra for VoLTE? Nah. 4G is good enough for me. And considering the cost to the telcos... something doesn't add up. As far as I can see, they simply can't add on enough consumer fluff to pay for what is a massive investment.
The only answer I can see, the only untapped market, is IOT. And my only example is some car company made a deal with some big US telco to do all of their car originating data, and I think that is where the telcos will make their money.
But again, I am coming from the software side. 5G to me isn't about radio or and of that, it is about telcos moving their compute to the edge, and virtualising their entire infrastructure to allow them to roll out new services in weeks instead of years. And I am at least a year out of date ;)