There are successful equipment manufacturers; my favorite is Ubiquiti Networks. They're competing (directly) with Cisco's wifi business. Cisco's carrier-switch side is, in turn, competing with Alcatel-Lucent, the traditional supplier to AT&T.
As for carriers, the important pieces to the equation are bandwidth, distance, and coverage. In the United States you need all three, because we have a lot of people spread out over a lot of space. There are 400 million people in this country. Traditionally each of them is connected to a telephone central office, where the equipment lives.
You see all these little dots? http://en.wikipedia.org/wiki/File:Central_Office_Locations.p... Each one of those is a building with at least a few hundred (and sometimes up to a hundred thousand) subscribers connected to it. Each building is a data center of sorts, sitting on land that had to be bought, containing equipment that needs maintenance occasionally, and getting power bills that need to be paid on time.
Many of these dots, particularly the isolated ones, are served by only one carrier, AT&T, who then leases out bandwidth to their competitors because the federal government requires them to do so as a condition of the antitrust settlement in 1984.
Universal service means you're connecting every last one of these dots to each other dot, and with sufficient bandwidth to make that work. Wifi doesn't have that kind of bandwidth, and it's not efficient anyway. The best way we currently have to do this work is optical fiber, strung between poles or plowed into the ground.
I'm not saying it shouldn't be done. I'm just saying that any "startup" who goes at it will need to be well funded.