Thursday, January 29, 2015

AWS-3 Auction Ends: Prices Arguably are at "Bubble" Levels

It isn’t yet clear precisely how much the price of one MHz-pop of mobile spectrum in the U.S. market has changed, now that the AWS-3 spectrum auction has concluded.

But it is fair to say the amounts bid likely will represent prices that are as much as 250 percent higher than the last auction of AWS-2 mobile spectrum.

Few, if any, expected the spectrum prices to be bid at such levels. Even in the early bidding rounds, prices grew exponentially.

It is hard to say how the sharp rise in prices might affect future auctions, such as the 600-MHz TV spectrum auctions scheduled for 2016.

Some might argue the AWS-3 prices were so high because many bidders think the 600-MHz auctions now are too uncertain, both in terms of spectrum that might be released, and the uniformity of released spectrum in the major markets where the need for spectrum is greatest.

AT&T and Verizon, the expected big winners of AWS-3 spectrum, reasonably would be working under the assumption they cannot gain big chunks of new spectrum by buying other mobile service providers, as regulatory authorities clearly have signaled they will not allow that to happen.

So even if other entities (Comcast, Sprint, T-Mobile US and others) might have other options, neither AT&T nor Verizon can hope to acquire lots of new spectrum by buying another large U.S. mobile service provider.

There is the potential chance to acquire spectrum, but no other assets, from Dish Network, which, with the recent prices bid for AWS-3 spectrum, might conclude it can sell its spectrum and avoid getting into the mobile business at all. Some have speculated that has been Dish Network’s possible objective all along.

One might note that the AWS-3 prices are higher than anything seen since the Internet and telecom bubble days leading up to 2000. That will hardly be reassuring for many observers.

No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...