Recent FCC statistical reports suggest that U.S. broadband development is headed in some surprising directions. Wireline-based broadband may become an adjunct to subscription TV services while wireless broadband should become the dominant Internet access vehicle but with lower than expected data rates.
The FCC's National Broadband Plan provides a benchmark for modern broadband. It observes that legacy Web browsing, email, and YouTube quality video streaming require a 1 Mbps download data rate, while these applications plus "near-real-time" media require a 4 Mbps data rate. Near-real-time media generally refers to video embedded in websites and the current generation of video streaming services. Using these requirements the FCC sets broadband service with at least 3 Mbps download and 768 Kbps upload speeds as a modern U.S. broadband baseline.
The FCC baseline is very low compared to worldwide measured speeds. Ookla, a Seattle-based broadband testing service, claims to perform 50 million broadband speed tests a month. It estimates average global household broadband speeds at 9.6 Mbps download and 3.6 Mbps upload. The U.S. ranks 33rd on Ookla's download speed rate list while Lithuania is first.
The latest FCC broadband development report (year-end 2010) finds that 37 percent of U.S. households have broadband service that is advertised to have greater than 3 Mbps download and 768 Kbps upload speeds. Wireline service (DSL, cable modem, and FTTP) accounts for 87 percent of these connections while 13 percent are mobile wireless broadband. Note that this report uses advertised data rates and that a year ago the FCC found that actual rates were about one-half the advertized rates.
Cable modem service accounts for 64 percent of the residential connections that are at or above the FCC baseline data rate. DSL and mobile wireless each have 13 percent share while Fiber to the Premises (FTTP) has 9 percent share. Growth rates over the last six months of 2010 were 4.4 percent for cable modems, 15.5 percent for DSL, 91 percent for mobile wireless, and 12.5 percent for FTTP.
Future wireline broadband development depends primarily on the development of FTTP connections as telcos migrate from DSL to FTTP. AT&T (NYSE: T) reported an annual rate of growth of 38 percent in its U-verse service in its most recent quarter while Verizon (NYSE: VZ) reported 12 percent for its FiOS service. Some of this growth is due to cannibalization of older DSL services. Though recent FTTP growth rates are strong the companies have a long way to go to reach the scale of their residential voice businesses. The installed base of U-verse and FiOS are each about 20 percent of each company's residential voice connection installed base.
At the end of last year only 10 percent of mobile broadband connections had data rates in excess of the FCC baseline. Most of these connections used 3G technologies where the advertised rates are about 1 Mbps download and 0.5 Mbps upload. 4G LTE offers from AT&T and Verizon are about 10 times faster. Verizon, for example, advertises 4G LTE as 5-12 Mbps download and 2-5 Mbps upload. While greatly improved over 3G data rates 4G LTE is still well below basic cable modem broadband which advertises an 18 Mbps data rate. Original 4G proposals called for multichannel deployments with data rates in excess of 100 Mbps. This would put 4G at parity with the cable industry's DOCSIS 3.0 technology. Multichannel 4G does not seem to be on most operators' roadmaps at this time.
I expect to see the development of a two-tier broadband market. Wireline operators will slowly move to FTTP solutions with data rates somewhat higher than current levels. However, I don't expect the U.S. to become a world leader as measured by the speed of our broadband connections. The broadband operator's dilemma is that subscription TV revenues are needed to cover somewhat more than one-half of the FTTP plant investment. A broadband offer with data rates equal to those of the global leaders is likely to undercut the operator's subscription TV services.
Market needs and investment requirements also seem likely to inhibit wireless operators' move to multichannel 4G. First, multichannel base stations and backhaul facilities will require substantial incremental investments on top of those being made now to migrate to LTE. Second, the necessary additional radio spectrum may not be available; and third, smartphones and their popular apps perform very well with the current generation of single channel LTE (or advanced 3G) technology.
Consequently, I expect most consumer Internet activity to be dominated by wireless broadband with performance fairly close to the FCC baseline, while something less than one-half of U.S. households will use high-end wireline services as part of triple play service packages.