About two points, there can't be much doubt. There isn't a lot of competition in the U.S. broadband market and U.S. broadband speeds are relatively low in international terms and—crucially—much slower than is feasible with current technology. Verizon FIOS and AT&T U-Verse exist, they work, and you can even buy them in some places. But not many places.
This leads recovering libertarian Tim Lee to conclude that U.S. broadband policy isn't working, citing in particular his disappointment that "in 2010 Verizon announced that it would stop installing fiber without reaching some of its most important markets, including Baltimore, downtown Boston, and my own apartment in Philadelphia" and that none of the other Baby Bells "have any further plans to run fiber optic cables to peoples' homes." His proposed policy response is competition-promoting regulation and there's an effort to square that with libertarian commitments via an interesting discussion of Hayekian vs Schumpeterian theories of innovation.
But in this particular case, we're not really talking about innovation at all. What's galling—to those of us who are galled by it—about America's broadband infrastructure is that we don't need more innovation. The relevant technology has been invented. What we need are more holes in the ground so that we can fill them with fiber optic cables. This is an expensive proposition, and the basic problem here is that when Verizon dipped its toes into making the investment it turned out that customers didn't want to pay for it. People who write about tech policy on the internet for a living place an unusually high subject value on high-speed internet access, so the resulting situation is frustrating for us. But the problem is really with the customers rather than the companies. There simply aren't enough people who want to pay a premium price for premium service.
One way to change that would be to broadly increase household income. This is the kind of situation in which economic inequality spurs technological stagnation. A household earning 20 times the national average income isn't going to buy 20 broadband connections. But if the median household were richer, it might be more willing to pay up for better internet service.
Another would be industrial policy. It seems plausible to say that whether or not households want to switch to a higher point on the broadband price/quality curve that it's in the countries long-term interests to encourage them to do so. Places like South Korea where the fiber is blazingly fast have clearly made this policy choice. And here in the United States we've made a number of policy choices aimed at getting people to build and purchase larger houses than they would in an unsubsidized market. Incentivizing broadband internet investment seems more sensible than incentivizing large houses, and if we want to make that policy choice, there are any number of ways to make that happen either through direct subsidies or through regulations designed to create cross-subsidy. Right now, however, I think it's less that our broadband policy is failing than that for all the white papers there isn't much of a policy. Better technology exists, but building it out is expensive and most customers don't seem to want to buy it.