AWS-3 Auction Predictions

I published this via twitter a week or two ago, and now I’ve cleaned it up and am publishing it here. In general terms, I expect Verizon, AT&T and T-Mobile to each own about 40-50MHz in the AWS-1/3 combined spectrum “superband” after the auction closes. This, after some deals between the three of them to reorganize the spectrum so its contiguous, will allow each of them to deliver end-user speeds of 15-25Mbps in areas of good coverage.

[Update 3/31/2014: Updated to reflect FCC’s band plan for AWS-3, also tweaked the results slightly to accommodate the new band plan]

Combined with lower band 600 and 700 MHz spectrum, along with refarming of PCS 1900MHz spectrum over the next 5 years, LTE coverage should be greatly improved in terms of speed and capacity.

What is left to determine is how end-user usage patterns will change when they have access to higher speeds – will they need to move up from a 10GB bucket to a 15GB bucket for their family of four lines, or will users keep their habits in check and the carriers are left with an overbuilt network because people aren’t streaming video for fear of data overages?

How a Sprint/T-Mobile merger should happen

Disclaimer: let me say I’m against the merger today. However if it were to happen, this is how it should happen.

First, T-Mobile should be allowed to succeed for fail on its own. Its fourth quarter numbers weren’t great. If T-Mo can turn it around on their own and become profitable in the US and they don’t need to be bought out, then great. If they cant, and we get to the point where the US market has proven that it cant sustain four profitable, healthy carrier networks, then T-Mobile and Sprint should be allowed to merge. But this means giving T-Mobile time to fix and market themselves.

In the same vein, Sprint should be allowed to fix themselves and open themselves to other opportunities. Masayoshi Son said recently about offering wire-line replacement at speeds of 200Mbit/s using 2.5GHz BRS/EBS spectrum.

I’m guessing we wont be able to make this judgment on whether current actions will turn around both companies until late 2016 or 2017. So any merger shouldn’t be allowed until then.

By then, what would a Sprint/T-Mobile merger look like? Probably a lot easier than what it would take in 2014 or 2015. Most notably, omni-band phones would allow Sprint to push most of its consumer and business customers over to T-Mobile’s faster GSM/LTE network rather easily and quickly.

T-Mobile will have some 700-A spectrum assets to leverage for better LTE coverage, as well as upcoming 600MHz band assets as well. They’ll have likely rounded out their AWS spectrum with spectrum from the AWS-3 auction to ensure 20+20 LTE in the top 50 markets (they only need 10MHz in many markets, 20MHz in a few markets). There will also be opportunities post-auction to rebalance spectrum assets from market to market, as Verizon, T-Mobile and AT&T look to rebalance their spectrum block between markets such that they own continuous blocks of spectrum – in most cases, Verizon will occupy the lower 40MHz (A&B), T-Mobile the middle 40/50 (C, D, E and F), and AT&T the upper 40/50 (future blocks in the AWS-3 auction).

Sprint will offer supplemental LTE with their 2.5GHz small cells to provide higher capacity and speeds in dense metro areas. Sprint’s legacy CDMA will have to stick around a long while for M2M customers, but they could start refarming CDMA capacity in the PCS band for additional LTE.

From a technology standpoint, waiting another two years down the road makes it easier. But will the owners of T-Mobile and Sprint want to wait that long? I think they’d like to go right now. Its up to the FCC and DOJ to make sure that we have four viable carriers in the marketplace until the marketplace proves definitively that its not going to work.