The problem is that there is a theoretical limit to how wireless infrastructure can be improved. Sure, new generations of wireless tech. are improving the bandwidth available every few years. But RF frequency bands are a limited resource and sooner or later Shannon-Hartley's theorem [1] will dictate that the only way to improve bandwidth will be to increase power levels received at user's antennae (either by increasing effective transmission power or clustering cells more densely). But doing this will deliver ever diminishing (logarithmic) gains and an increasing public health risk associated with increased RF levels.
Couple that with the fact that wireless is going to be the Internet in the coming decades (for many youngsters it already is) and you are faced with a problem of how to allocate things fairly.
Wired Internet resources on the other hand can be upgraded effectively forever. If the demand is there we can keep laying new fibre whilst maximising the existing lines with new modulation techniques etc. The same cannot be said of wireless technology.
It is precisely because of these major contention issues that wireless spectrum should be protected under net neutrality regulations, its need of regulation is in fact greater than that of wired Internet. When there is an approaching hard limit to bandwidth available why would we want to allow the parties with the biggest wallets dominate the traffic? It is precisely these sorts of situations that governments are for, without them monopolies will naturally form at the expense of the populace.
Imagine if governments around the world said, let's solve (or rather ignore) the problem of car traffic congestion by allowing employers to bid for priority access to the roads on behalf of their employees, and businesses to bid on behalf of their customers, even though everyone has already paid their own car tax, fuel etc.
[1] http://en.wikipedia.org/wiki/Shannon%E2%80%93Hartley_theorem