It's the same principle that led municipalities to grant taxicab monopolies, of the kind that stifle companies like Uber today. The idea is that in a free market, providers would only focus on the profitable parts of the city, leaving the lower-income areas without service.
Universal access is why it's impossible to "disrupt" the market for local internet. A new entrant can't just come into the market and pick off the most desirable customers or the ones that are cheapest to service. They have to be prepared to service everyone, even many customers who can only be served at a loss, in order to be allowed to operate at all. Additional regulation in this space isn't going to eliminate the underlying problem, it will just make infrastructure construction a more unattractive business and decrease investment.
There is no free lunch. You can have: neutrality, universal access, or a mostly privately-funded telecom infrastructure, but you only get to pick two. If you think universal access is important, and net neutrality is important, you have to be willing to publicly subsidize the construction of telecom infrastructure.
PS. Some folks have mentioned BT OpenReach, but it's important to understand what did and did not happen there. First, BT was originally a government-owned corporation before it was privatized, so the government was in a position to set the terms of the privatization. Second, the government used a fairly generous "RPI - X" price cap to certain of BT's services. Between monopoly pricing power and a relatively generous price cap, BT made very healthy returns on investment. See: http://www.academia.edu/3399930/LESSONS_OF_PRIVATISATION_IN_... (p. 28-9).
In the U.S., unbundled DSL was a failure. First, it got caught up in litigation because the infrastructure was never public at any point. Second, it was an almost pure losing proposition for the telecom companies. The FCC mandated a not-very-generous cost-based price control (TELRIC), which made further investments in DSL infrastructure unattractive.