Template Tools

We've moved

Everything since June 1 is at  fastnet.news. This is just the archive from before June, 2015

Fiber Economics - Quick and Dirty
Written by Dave Burstein   
Saturday, 11 August 2012 19:04

Milo_building_KC_and_former_partner_John Muleta~$1,000 per home passed to build, $5/month to operate. Assuming they get a decent take rate and meet their construction budget, Google will be profitable charging $70/month in Kansas City. But the payoff requires thinking more than three to five years, unlike most corporations. Here’s the data behind that conclusion.

Construction costs, typical city or suburb
Bell Alliant in Western Canada has now passed over half a million homes with fiber home, the largest deployment in North America after Verizon. Their latest financial report showed capex of less than $500 per home passed. Verizon reported costs fell below $700/home passed several years ago and headed to $600. Add the cost of actually installing a large fraction of those homes, and your cost per home passed by the network comes closer to $1,000. Installing each home at Verizon added $500-600. Digging lawns and drilling holes into the homes is labor intensive.
    That includes equipment whose price is rapidly dropping. Early Verizon gear cost $300-400/home, but today they are probably paying half that. Very large fiber builds in China are paying less than $100/home. Google has designed their own gear, GigaOm reports, and is going to the least expensive sources for components and assembly. They should be achieving something close to the large volume China price.
   Fiber itself cost $20K or so per mile, mostly labor. Rural areas with fewer homes per mile can be brutally expensive. often $3-5K. Undergrounding the fiber also adds costs.
Operating costs:
Bandwidth isn’t free, but it’s darn cheap. A moderately sized carrier pays less than $1/month/customer http://bit.ly/N3kVKS. Google pays much less, while small and rural carriers sometimes pay much more. Bandwidth use goes up only modestly with higher speeds; the industry rule of thumb is about 1/3rd more usage if you give the customer higher speeds. A Netflix movie will come in at 2-4 megabits whether you have 12 megabit service or a gigabit. You don’t get more email because your connection is faster.
Line maintenance and repair should be extremely modest. Google’s fiber architecture has no active components in the field. Verizon’s expectation was that fiber would bring down maintenance by 80%. Their experience was very different, as it took several years to get the customer installs reliable. Five years later, most of those problems should be solved. Google in particular has reduced network complexity to keep opex down.
Support costs for broadband have gone down dramatically. New broadband customers typically make as many support calls in the first 90 days as they will later in a year. Most customers today are well down that learning curve.
Power is a factor, but current equipment keeps that well below $1/month/customer

The crucial factor of take rates:
The cost per home passed is very different than the cost per paying subscriber. The majority of the cost is building the network, not the home hookup. If you have only a 20% take rate, the per home passed figure has to be multiplied by 5. At 33%, that drops to a multiplier of 3. At the likely Australian 70-80%, only 1.5. (In Australia, the NBN is paying Telstra and others not to compete and ensuring something close to a monopoly.)
    At $600/home passed, it costs $1-$1.5K/subscriber in Australia, $2-3K/subscriber with a 33% take rate (Google’s hope for Kansas City) and $4-5K/sub if you only get 20% of the market.
     $1K/sub is $16/month over 5 years, little more than the cost of unbundling. $3K - likely Google expectation in KC - it’s $50/month. At $5K it’s $80/month. Even at $70/month, you have to be thinking 10 years for a good payback.
      Governments paying 4% on long term bonds can think 10 years out, especially with the external returns to having a better Internet. Private telcos rarely think beyond 3-5 years, one reason nearly no telcos are willing to invest in fiber unless they have strong competition or a government push. VDSL from field cabinets is now delivering 50-100 meg in many locations and has become the near exclusive choice of British Telecom, AT&T, and Deutsche Telekom for the faster payback.
  These numbers vary enormously depending on neighborhood conditions and the efficiency of construction. Burlington Vermont’s muni fiber came in excessive construction costs and now is a scandalous bankruptcy. Australia’s NBN budgeted $5,000/home and is already over budget. Free in Paris is very disappointed at the hookup costs per home.

   Milo Medin of Google built America's cable modem networks 15 years ago and knows how to do this right, so I'm optimistic for Kansas City.