I think Edward Wyatt buried the lead in his article on U.S. broadband speeds and broadband adoption. Here's the 16th graf of the piece:
San Antonio’s power company has a largely unused fiber-optic network that local government offices have been using for high-speed Internet service for years, but a Texas law prevents the city from using the network to give low-cost service to consumers.
To clarify something Wyatt leaves out here, the power company in question is CPS Energy, which is wholly owned by the city of San Antonio. CPS provides service to Bexar County outside the city limits and to portions of the seven counties that surround the city. So it's an interesting exercise in a form of market socialism that's rare in the United States.
And CPS Energy already has the fiber. So it's not a question of should a municipally owned utility spend money on building a fiber-optic network. It's a question of given that the municipally owned utility already has a fiber-optic network, shouldn't it do something with it? The Texas state Legislature, doing the bidding of local telecom firms, says no. It can't.
The larger question Wyatt raises in his piece of whether the United States as a whole is lagging behind the rest of the world in Internet speed gets a little bit tricky to ponder. The difference of geographic size between the United States and Latvia means that comparing the two isn't particularly enlightening. But the flip side of being a big country is that there's lots of room for localized policy failures. I'm not sure that every city in America should build a municipal fiber service, but it's a perfectly plausible approach to the problem and it's clear that some experimentation is useful. And it's utterly clear that a city that already has a municipally owned utility that's already laid the fiber ought to press forward with serving the mass public.