Engineers know this because they stand on the shoulders of giants. Policy people are too often blinded by the fog of interests. With just the spectrum already allocated, AT&T and Verizon could easily triple their network capacity without breaking the capital budget. The technology is obvious but the change in business models is unlikely. Sprint has remarkable holdings around 2.5GHz. Dish spectrum is totally unused.
There's enough available spectrum to build at least three Verizon-sized networks. The currently allocated but unused spectrum is allowing both AT&T and Verizon to double their wireless capabilities. I wrote this article because <yet another generally very knowledgeable> person just left that doubling out of his predictions. It's commonly overlooked.
My estimate is that spectrum, advanced technology and investment can easily provide a 25x gain many without blowing out capex budgets - if policy is smarter. The increased costs should be no more than 2% of sales and probably much less.