Chorus is to launch discussions with interested parties
regarding re-working the UFB fibre plans and prices.
The entry level 30Mbit/s down, 10Mbit/s up plan costs $37.50
per line per month (increasing to $42.50 by 2019) at a wholesale level.
That doesn’t include national backhaul, international
backhaul or any of the other stuff – it’s just the Chorus bit, so don’t expect
to pay $50/month for fibre services any time soon.
What’s interesting is the range of new products – from a
50/20 plan through to 200/200.
Most of the current plans have a committed information rate
(CIR) of 2.5Mbit/s – that is, if your line is utterly saturated with use, that’s
the minimum you’ll get.
That sounds awful, until you consider the CIR on copper
wholesale, which at the basic level gives you 45kbit/s, which is dial-up speed.
You’ll see the biggest mover is in the business space with a
doubling of the high-end plan’s speed from 100/100 for $175 to 200/200. Now
that’s a plan to get excited about.
But what will the costs be like for the retail service
providers? Don’t forget, if you as a customer want to see these speeds
throughout your entire network, your ISP is going to have to buy a lot more
An anonymous ISP source has done some numbers for TUANZ on
the Gigatown promotion that Chorus is running. Gigatown, you’ll remember, is
the plan to offer 1000Mbit/s service for the 30/10 price. However, Chorus is
unable to sell directly to users, so it has to bring in a retail ISP to do that
side of things.
Let’s assume a town of 40,000 people is chosen for Gigatown.
Chorus will provide the ports at entry-level price, so 40,000 customers x
$37.50 = $1.5M/month
Upfront costs are not cheap:
First you’ll need equipment at the exchange to handle
the traffic. A fully-loaded chassis will
handle around 7,000 ports, so that’s six
chassis and 100x10Gbit/s backhaul. Roughly a $6 million cost for exchange
The absolute minimum price for a box that can handle a
10Gbit/s backhaul is $1200 for a Mikrotik CCR, so the RSP has at least $120,000
capex spend to get in the game.
Then there are the recurring monthly charges:
Either co-locate the equipment at the exchange or get
100x10Gbps backhaul to premises, either way it’s roughly $30,000 per month.
The RSP needs to provision bandwidth for the
customer. Let’s assume a generous $1/Mbit/s
for national, provision – 100Gbit/s will cost $100,000 each month.
Then there’s international capacity at $17/Mbit/s
equivalent (ISPs don’t buy international bandwidth in these terms so this is a
bit of a translation), so to provision 20Gbit/s assume a cost of $340,000 a
“With no other costs, the RSP would be at $2M/month,
with over $1.5M of that going to Chorus, assuming 100% penetration.”
At today’s rate of about 5% penetration (current UFB
stats), your costs would come down but so too would your earning potential.
Five percent uptake means 2,000 households, so only
one chassis is needed, 14x 10Gbit/s backhaul and an upfront cost of only $1
million capex at the exchange, and $20K capex for the RSP.
Backhaul or colocation costs would come in at $10,000
a month, national backhaul would be
another $20,000/month and then your international would add another
Those 2,000 customers would earn $75,000 a month
(assuming no calls to the call centre etc) but the costs per month would be around
$205,000, so you’d have to bill customers around $102.50 a month to break even –
no profit for you.
I’d pay $102.50 a month for gigabit speeds without a
second thought, but would you get 100% uptake at that price? I don’t know.
What does all this mean? If the sums are right (and do
let me know if you think they don’t stack up and we can discuss tweaking the
model) then selling UFB is going to be a big chore for RSPs. They’re going to
need faster speeds (but that comes at a cost) and will need to give customers a
reason to move (content, for example, which also isn’t cheap) and they’re going
to need help with the marketing to get the general population excited about it.
These really are interesting times.