My examples were monopoly situations but the lack of "upgrades" were not because it was a monopoly
We'll never know, at least in those examples, because you chose to use monopoly examples.
"100% of peak possible demand".
Ok, let us get this out of the picture because this is mentioned a lot in your post. Chas mentioned that internet capacity is deliberately under-provisioned. For this to be false, the ISPs do not have to satisfy 100% of peak possible demand - this is completely your invention. Even if the ISPs do not invest reasonably in improving network infrastructure, knowing well as every 10 year old knows for last 10 years that internet is a growing business, the under-provisioning is DELIBERATE. It is not just because it is the failure to satisfy 100% peak possible demand.
Hope this is understood - I am ignoring rest of your references to your absurd 100% peak possible demand which no one is talking about except you.
Running 100 long distance trunks to serve a 100 person central office would be "100% peak possible demand", but it would be outrageously expensive.
People talking long distances, while a growing business, is not as much a growing business as internet services. Taxis in new urban areas, or newly prosperous areas, are a much better example and their the taxi operators at times do have enough provisions for peak demand because peak is expected to grow. This also shows that lack of monopoly made the taxi operators upgrade the fleet. You mention that ISP's reluctance in upgrading is not due to monopolies without giving any reason - except the absurd interpretation of upgrade to mean 100% peak possible capacity.
The problem here is the fact that ISP business is a situation where customers don't have a real choice.
Even were that true, it would be irrelevant.
Taxi example proves it is relevant. You do not give any reason why it is not.
And at times the demand exceeds supply. They don't buy more taxis to cover that, they can't afford it.
They do in fast growing markets.
So you're saying that Netflix has to pay to increase the capacity of the delivery system (I was Netflix in the analogy, in case you missed it.)
1. If you were Netflix in that analogy, you don't understand the situation at all. You said "I buy 100 Mbps link", and your situation had YOUR link become the bottleneck in serving 2 100 Mbps customers. Netflix situation is not like that.
2. It is not impossible for a content provider to profitably provide for increasing their customer's internet connectivity - see Google/Amazon's plans/dreams/bluffs for providing drone wifi-access points in non-connected areas, e.g. Africa. Google bid pi billion dollars to ensure "net neutrality", and is investing in Google Fiber, probably at a loss. All this had Google "invest" around less than a dollar per potential customer. If Google had much higher margin customers, they could have invested the 100 s of dollars per customer that upgrading 100 Mbps to 200Mbps requires.
Coming to your example, I didn't mention that, because typically it takes a lot more than 2 customers to be able to do such things - probably billion customers. But if the 2 customers give you enough margin, sure, why not. Netflix has nowhere close to billion customers, nor a margin enough to do that.
Earlier you said that customers don't have a choice. And now you say I would choose a different provider.
Exactly. Which is why you are having to ask who pays for upgrade. If you had a choice, it would have been upgraded by now.