Fast Net News, DSL Prime, DOCSIS Report

With word from Cox they intend to do two-thirds of their territory by 2010, I've raised the estimate for 2010 to 75 million U.S homes able to get 50 megabit service DOCSIS 3.0 by 2010-2011, more than half the country. At least half of the remainder will be covered in 2012-2013. 20+ million are ready today: 17M at Comcast, and Cablevision is ready to turn on 5M more. I've also raised my estimates of Time Warner and other companies after Saul Hansell in the NY Times reported DOSCIS upgrade costs as low as $20.

Few outside industry circles has thought this through, although it's a technological breakthrough that will change the Internet and should revolutionize policy thinking.

Since Comcast covers almost half the country and Brian Roberts has promised to serve it all by 2010, even with delays it's almost certainly 60% by 2011. The other numbers mostly are based on official company presentations to Wall Street, where CEOs try hard not to lie.

DOCSIS 3.0 is deploying amazingly quickly because the $100/home cost is so cheap and the perfomance is proving outstanding. We won't know until nodes start getting crowded, but 50 meg 95+% of the time is highly likely. For $100/ most homes, every cableco without financial problems is upgrading very quickly. All of Cablevision and J:COM is upgraded, most of Numericable in France, and much of Virgin in Britain.

3.0 is already deployed to close to 20M U.S. homes. Cablevision told the street they are ready with 5M homes they will turn on in within months. Comcast is 17M million, and some significant number at Cox and others. These numbers are direct from the companies, usually on financial presentations. I've crosschecked against shipment by the vendors, and that's about right.

3.0 will almost certainly be deployed to 60-80M U.S. homes within three years. That's based on 40M Comcast says for 2010, 5M Cablevision, and somewhere between 20 and 40% at Time Warner , Cox (unofficial source, but I believe more likely to be close to Comcast), etc. with very like from Charter and some of the smaller ones.

3.0 is highly likely to be deployed to 75-85% of U.S. homes in 2012-2013.
The DOCSIS deployment timetable makes sense when you consider the very low cost of the upgrade, placed by Tom Rutledge of Cablevision at less than $100 in 2008 and presumably dropping. Comcast believes the figure is closer to $50, according to Wired.

3.0 is 160/120 shared minimum, which I believe (but can't prove yet ) is 50 up, 50 down over 95% of the time in likely U.S. deployments. More recent modem chips are 320/160, Cisco has shown 400, and the spec is designed to go to a gig symmetric. Among other things, that means it would be easy to hit any arbitrary 50 or 100 meg speed test in 2011-2014, simply as Moore Law's improves the chips and SDV/analog shutdown makes the channels available.

Time Warner when I sent this to them for factchecking made a point of neither confirming nor denying my estimates. My primary source is CEO Glenn Britt, who told investors they have the technology and will roll it out in competitive markets. They face Verizon FIOS in New York City, which is designed for 200 megabits and will mostly be installed by 2010-2011 under the Verizon franchise agreement. I don't have a number for the Time Warner Cable/Verizon overlap, but I believe that alone will drive 20% in 2009 or 2010. In addition, I don't believe TWC will be more than three years behind Comcast, which is set for 100% in 2010. That leaves me comfortable with a 90% estimate for 2013 based on the the underlying business issues, the fact that the upgrade is so cheap as to be a no brainer, and the uproar likely from Government if Time Warner services are so much slower than other cablecos.

Cox has now announced they plan two-third coverage by 2010. They added they believed this was a good time to invest in upgrading the cable business. On the other hand, they are spending $2B+ to become a major wireless player, a story worth telling at length. Cox bought the spectrum and have very rapid deployment plans for wireless. The vendors, Huawei in particular, are offering extraordinary prices and possible vendor financing on the wireless gear, which will keep the cost down. They have a tough decision about femtocells, which would require higher capital spending up front but may be necessary to meet AT&T's planned wireless cloud of 10 million femtos across their half of the United States. Cox will mostly sell in district as a quadplay, keeping customer acquisition costs down. The cd-quality voice codecs in the DOCSIS 3.0 standard has been of interest to them for several years. Deploying high bit rate voice telephony in market would give them a major technical advantage over the telcos. They have enough concentration in many markets to make it work, but have announced no deployment plans. They've always prided themselves on having a technically excellent network that provides better services to their customers than most other cablecos.  I think they are more likely to be above rather than below the estimate, but there's little public information.

Charter rescheduled debt but now have filed Chapter 11. They are far behind the other large cablecos in deploying phone and high speeds so far, but they have been early technical leaders in DOCSIS 3.0. They may surprise people. The new FCC doesn't want consumers to suffer because the company that serves them is overleveraged. The loans in the stimulus package are one way to deal with this, as well as cash short companies like Suddenlink and Time Warner.

Cablevision and Jimmy Dolan deserves kudos for being the fastest to deploy in North America.

A wild card in the new FCC under Obama. The transition team decided to kill a large subsidy for telco broadband deployment partly because they knew DOCSIS would reach most of the country quickly. They may prod cablecos who fall behind in deployment. Jules, unlike Martin and Powell, is very good at looking CEOs in the face and making sure they follow his lead.

As always, all help from readers appreciated.