The FCC’s AWS-3 auction has finally ended, with a total of a bit over $41 billion (after discounts) being spent nationwide for 65 MHz of spectrum. The auction was dominated by three bidders: Verizon, AT&T, and Dish Networks.

Verizon and AT&T were clearly aiming to bolster their spectrum holdings in anticipation of continued rapid growth in demand for broadband services. The Dish strategy is not as obvious. They spent a lot of money to gobble up paired spectrum in big markets like New York City and Chicago, and for a pretty much complete nationwide footprint in unpaired spectrum. Some analysts believe that Dish intends to lease spectrum to Verizon in key areas, but unless they have a deal already worked out the $10 billion or so they spent in the auction would seem to be an awfully big gamble.

What I don’t expect to see is Dish attempting to leverage their profitable satellite TV broadcast business to deploy some sort of satellite-based (or hybrid satellite/terrestrial) Internet service along the lines of the failed LightSquared effort a few years back. With some very limited exceptions, providing broadband connectivity via satellite is simply unworkable both technically and economically. The exceptions, which companies like Hughes Networks have marginally exploited, are for providing modest Internet speeds to fixed terminals in truly isolated areas where there are no alternatives.

Having spent a total of $18.2 billion and $10.4 billion respectively, AT&T and Verizon are likely more than a little bit nervous about their AWS-3 investments. Yes, they added key chunks of spectrum to their holdings, but at a very high cost and at a time when broadband revenues, in terms of dollars per MHz of exploited spectrum, are at best flat. And the revenue picture is probably going to get worse in the short term as the current price war among nationwide carriers heats up.

So far, the industry’s only answer to surging demand for wireless broadband service has been increasing the amount of spectrum available to broadband networks. Obviously, that’s not a sustainable strategy. For example, Verizon already controls over 100 MHz of spectrum in almost all of the biggest US markets, and yet they struggle mightily to keep up with demand in high traffic areas. Does anyone really believe that adding 10 or 20 MHz more will solve their long-term capacity problems?

If there is a glimmer of silver lining in all of this it’s that today’s LTE networks are still operating at very poor levels of efficiency. In particular, engineering LTE networks for increased usage density is being hampered by a lack of good interference management tools. What this means is that a significant effort aimed at network optimization is likely to result in substantial gains. I suspect that if even a small fraction of the billions spent in the AWS-3 auction had been used instead for development of sophisticated interference management tools for LTE, the result would have been a far greater increase in wireless broadband capacity.

I certainly can’t fault the strategy of AT&T and Verizon, and Dish Networks for that matter, in spending big for AWS-3 licenses. But in the long term the industry needs to do much more to address the need for capacity than simply bolting on more spectrum. Resources, both financial and human, need to be prioritized accordingly.