Our Balkanized Broadband Future

Andrew Cohill has made some apt observations regarding a likely future of broadband in the United States. The thesis is that a few providers can effectively disrupt the likelihood of an entire community getting next-generation services by locking up key customers. And I agree.
But today, the market for bandwidth continues to grow along a nice smooth curve, with the demand doubling every two years, and we have fifteen years of data to back this up. While the incumbents are busy trying to convince us they can meet this demand with 1950s copper cable plant, smaller telecom firms are busy spreading bits of fiber through communities to cherry pick the more profitable business customers. These companies tend to have no interest in full fiber build outs, and instead just want to lock up a portion of the local business market.
Some [not Cohill] have argued that when local governments stop overpaying for T1 lines and build their own networks to be fiscally responsible, incumbent telcos will be unable to continue investing there due to the reduced revenue. Of course, incumbent telcos have long ago ceased investing in these communities, so the proposition is off from the start. But even if it were true, it is an incredibly inefficient system (no matter how lucrative for the incumbent telcos). We need to actually start treating broadband as infrastructure (rather than simply talking about it as though it were infrastructure -- which most elected leaders seem to do). This means that when the community needs broadband, they are able to build it themselves and ensure the network will remain accountable to them in the future. The longer communities wait to build these networks, the more difficult a prospect it will be as private companies continue to pick off the high-revenue easy-to-serve subscribers.