I happened to talk to Someone In A Good Position to Know yesterday about the cost and scale of National's plan to bring fibreoptic cable to 75% of New Zealand homes within three five years of work commencing. The answer was not equivocal: at least six years from commencement at a cost of $6 billion.
By far the greater part of the cost (up to 80%) is in the works part: digging up the streets. The advantage of Telecom's cabinetisation initiative is that there is ducting from exchanges to nearly all cabinet locations. Past the cabinets, you're digging up every street in every suburb.
So I have a bit of a problem with Maurice Williamson's cheery assertion yesterday that everyone else is wrong, and we'd only be talking $2.5 billion, plus another $500 million to connect businesses. Making a total cost of $3 billion, which, handily, implies the private sector matching National's promised contribution dollar-for-dollar.
Williamson justified his estimate to Tom Pullar-Strecker by reeling off the cost-per-house for similar projects in such places as Amsterdam, Stockholm and Seoul. I'm not sure such comparisons are wise. Housing densities in Europe (let alone Seoul) are much higher than here. And, importantly, those places have access to a large pool of skilled, mobile labour.
We don't. Even for Telecom's cabinetisation project, Downer (the contractor to Telecom's operationally separated access network division, Chorus) is scrambling for every hand it can get. It's both training and bringing labour into the country. According to my advice, the 75%-of-homes project implies 200 fibre teams working five days a week for six years.
Cost? Let's assume that we're talking about passive optical networks, which employ a loop across which all users contend for bandwidth, rather than an active or "star" network on which every user has dedicated fibre.
We'll assume that not because it's better (it is considerably more limited), but because it's cheaper, and it's the way fledgling residential fibre is currently being built here. It's possible that splitters could be installed in the cabinets currently being installed by Telecom, with each household getting dedicated fibre from there, but you're talking about extra cost.
Williamson further justified National's costings with reference to Verizon's massive (18 million homes) FiOS project in the US, where he said cost-per-house is down to $NZ2158.
Woah. What Verizon is doing it pretty amazing. But the markets are nervous about returns and, so far as I can tell, Verizon is able to cut its costs because it's increasingly using existing coaxial cable to connect the passing fibre to the home. In many cases, it's also not undergrounding fibre, but stringing it from existing poles.
Further, Verizon uses its network exclusively and is relying on revenue of up to $US200 a month from each household and, probably, from TV operators. Returning a similar investment via wholesale revenue from an open-access network on which retail operators have to make their dollar too is another matter altogether. And remember, that number above is cost per house, not cost per subscriber. It would be perilous to assume 100% uptake.
To his credit, Williamson has given a much clearer steer as to the nature of the investment vehicle for National's project: it's the New Zealand Institute's FibreCo, by whatever name. Operators large and small would have the ability to take a stake alongside the government in the big company that builds the network.
We need to accelerate the deployment of fibre, initially to the node, and to the premises where there's demand. Where copper expires, naturally, you replace it with fibre. Ditto for new subdivisions (this is already happening, but I'd like to see someone other than Telecom doing it).
But I still don't buy the numbers on the Great Leap Forward. I think the building of fibre-optic cable infrastructure will perforce be piecemeal, incorporating copper technologies such as VDSL (if your cabinet is close, you might expect 50Mbit/s from that) and it will take a lot longer than National says.
I'm still a shoot-for-the-moon guy regarding fibre. But I'd prefer the rocket to take off from a reality-based platform. And I'll be interested to see what Labour announces in this week's Budget.
PS: Feel free to dig into the new OECD working party report on communications infrastructure and let me know what you think.