There are good reasons to expect market failure here. New players very rarely enter infrastructure markets with existing incumbents because having to compete with the incumbent in a natural monopoly market is not expected to be profitable.
And the "accidental duopoly" markets aren't going to do it. Both players try as hard as possible not to compete with each other because they both know that increasing speeds or lowering prices would only provoke the same response from the other and make them both less profitable.
So you're really asking whether a monopoly incumbent would do it, which doesn't have much to do with markets at all -- it turns on whether the cost is more than the increase in monopoly rent it would allow rather than whether the cost is more than the value. And the existing monopoly rent is already high because even slow internet is much more valuable than none.
> And if the government does build it, we should ditch the emotional appeals for FTTH and rely heavily on sensible, cost-effect point-to-point wireless.
FTTH is a one-time expense. The maintenance cost of underground fiber shouldn't be any higher than wireless, and then you don't need to worry about spectrum or wireless line of sight obstructions or interference or any of the other problems with wireless.