What will the National Broadband Network really cost?


This article is by Rod Tucker, Laureate Emeritus Professor, University of Melbourne. It originally appeared on The Conversation.

analysis Cost is a central issue in the ongoing debate about the best approach to building Australia’s National Broadband Network (NBN).

In 2013, the Coalition argued that Labor’s original all-fibre to the premises (FTTP) network could cost as much as A$94 billion. In the 2016 NBN Corporate plan the figure was revised down to A$74 billion to A$84 billion, while NBN Co’s multi technology mix (MTM), incorporating fibre to the node (FTTN) and upgraded hybrid fibre coax (HFC) was less costly, with a price tag of A$46 billion to A$56 billion.

Since the Coalition announced these numbers, Labor has said that, if elected, it will not go back to an all-FTTP network, but instead pursue a half-way option, in which the HFC component of the MTM is retained but FTTN, the slowest and most limited technology, is phased out.

It’s worth looking more closely at cost difference between FTTP and FTTN to see if the claimed A$84 billion to A$56 billion maximum cost comparison stacks up, and see where Labor’s new half-way solution sits.

Capital costs

The NBN 2016 Corporate Plan states that the average capital cost (capex) to connect a home or business to the NBN using FTTP is A$3,700. But the real cost for a FTTP connection is probably less than this.

The A$3,700 figure quoted by nbn co is based on old construction techniques that have been superseded in other parts of the world. The costs of rolling-out FTTP in New Zealand, for example, have been dropping steadily in recent years and will soon be A$2,900 per premises. For some reason, NBN Co has yet to acknowledge the lessons learned in New Zealand.

Let’s give NBN Co the benefit of the doubt and assume that the A$3,700 cost per premises for FTTP is correct. In comparison, the 2016 Corporate Plan states that the average capital cost for a FTTN connection is A$1,600, or A$2,100 less than FTTP.

For an upgraded hybrid-fibre-coax (HFC) connection the capital cost is A$1,100, or A$2,600 less. However, in light of recent revelations in a leaked document from NBN Co published by Fairfax indicating that it may be necessary to overbuild Optus’ HFC network, the savings offered by HFC will not be as good.

Using these numbers, it is easy to compare the capital costs of different networks.

At the end of construction, the MTM network will provide FTTN to 4.5 million premises and HFC to 4 million premises. Labor’s new approach is to replace as many as possible of these 4.5 million FTTN connections by FTTP. The maximum additional capital cost to do this would be 4.5 million times A$2,100, or A$9.5 billion. This figure corresponds to A$790 per premises averaged across all 12 million premises in Australia.

An all-FTTP network could be achieved by also replacing the HFC connections with FTTP. The total additional capital cost of this hypothetical all-FTTP network would be A$19.9 billion, or A$1,658 per premises averaged across all premises.

Peak funding

Of course, it is necessary to also consider operational expenditure (opex) – the cost of running the network – and revenues from the network. These factors all contribute to the peak funding figures in the 2016 Corporate Plan.

Peak funding is the maximum cash outlay required before cash flow becomes positive. Peak funding is a useful measure of cost because it is a direct measure of the cash outlay required. But it is not necessarily a good measure of the cost of the network infrastructure or a good measure of the net financial cost/benefit to the Australian taxpayer.

Operational expenditure is a major issue for the MTM network because of factors including the need for new software management systems, the additional costs of maintaining the degrading copper wires in the FTTN network, and the cost of the electricity required to power the FTTN nodes located in suburban streets. Importantly, an FTTP network would incur none of these costs.

In fact, the leaked NBN Co document mentioned earlier shows that the operational costs of FTTN network are 67% more than for FTTP, and the operational costs of HFC are 25% more. Over the lifetime of the network, this difference could amount to billions of dollars, greatly reducing the overall difference in costs between FTTN and FTTP.

Another factor that reduces the cost difference between the Coalition’s network and Labor’s new alternative is that a Labor’s FTTP/HFC network would be capable of generating higher revenues through the delivery of premium services that would not be achievable with a slower-speed FTTN network.

This is well documented by high-profile companies such as Ovum, which predicts FTTP services will drive the highest global growth rates for broadband revenues over the next five years, based on premium speeds of 100 Mbps and higher.


In light of all these factors, why is NBN Co’s cost estimate for a hypothetical FTTP network so large? The NBN Corporate plan provides no detailed information on its financial modelling, but it states that an all-FTTP network would take until 2026 to 2028 to complete.

If the timeframe was indeed as long as this, the revenue stream would be delayed. This could indeed lead to unrealistically large numbers for the peak funding cost of FTTP.

So where does the 2026 to 2028 timeframe come from? My guess is that NBN Co has simply extrapolated from the present rollout rate for FTTP, which has not increased much since 2013. One piece of supporting information comes from a Senate Estimates meeting, where NBN Co confirmed that its A$74-84 billion number was not for a “continued” FTTP network but for a “restart” from the current plan.

NBN Co waited until September this year to hire additional staff, increasing the number of employees from 3,400 to 4,500 to speed up the rollout of FTTN. If it had hired these additional staff in 2013 and focused on the FTTP rollout, the network could well have been completed by around 2020 or 2021.

By Rod Tucker, Laureate Emeritus Professor, University of Melbourne This article was originally published on The Conversation. Read the original article.


  1. Yep, Rod’s article really highlights that we are were we are today 100% due to Liberal interference.

    We really do need a Royal Commision into the Libs handling of the NBN if for no other reason than to show Joe Public how corrupt the Libs are and hopefully prevent any future govs (from either side) from committing the same type of politically motivated hatchet jobs on future National projects!

    • If the Liberals have their way, there won’t be future national projects.

      It would all be left up to the magical private sector who along with pixie dust and kisses, always have the public’s best interests at heart.

      • How many attempts did Howard have getting it off the ground, leaving it up to the private sector? Something around 16 or 17 I believe.

        Labor made one attempt. quickly coming to the conclusion it wouldn’t happen.

        For any hope, even being built by government, Telstra had to be taken out of the picture. Should have happened when Howard sold Telstra off.

        I have no idea of what we have now with MTM, coat of many colours with Telstra back in the picture.

        How one sorts the present mess out, will need a miracle maker.

  2. Another point to make that is often overlooked is that FTTP is capable of providing multiple services on the one fibre, e.g. FTTN is limited to Phone and Internet, while FTTP, just like HFC, can carry in addition to phone and internet, TV signals (Pay and FTA), that means that Foxtel and the FTA networks could cease their existing OTA broadcasts and instead lease space on the fibre (another income stream, with no added costs to the receiver, i.e. no broadband connection required), that would allow the government to make money from the sale of wireless spectrum (which would pay off part of the cost of rolling out FTTP). Currently FTA is limited to 35MHz of spectrum per market, but if fibre was used, it would be virtually limitless, allowing extra TV services, e.g. 4k or 3D and additional Services overall.

    IMO FTTP costs more than FTTN, but the benefits far outweigh the burden of extra upfront costs.

    • “IMO FTTP costs more than FTTN, but the benefits far outweigh the burden of extra upfront costs.”

      Which is why bean counters often miss the “B” in CBA.

      • @r0 bean coutbers typically good at both ‘B’ and ‘A’. Lets go to the opinion piece

        “But the real cost for a FTTP connection is probably less than this [$3700].”

        Probably? Actual CPP performance is on the next page in CP16. The $3700 is the best estimate using the new rollout techniques. Actuals using “old techniques” much higher.

        Seems onvious but Chorus’ ARs show their performance, NBNCo SRs shows theirs. Tucker might like to give Chorus’ FTTN CPP numbers (the vast majority of Chorus’ network; users and revenue) by wat of comparison. I asked Tucker’s esteemed college Gregory the same question; why NBNCo’s performance is so poor compared to NZ’s Chorus.

        CPP for FTTN is half of FTTH, a third for HFC. Tucker claims HFC savings won’t be as great given the Plan B leaked document. The new management increased contingency in their plans from 10% to 20%. Even if Plan B was utilised the additional costs supplied are well within contingency and tiny compared to the FTTH / HFC cost differential.

        The 4.5m FTTN figure includes 1m MDUs. Many will use FTTB. Given NBNCo’s old management failure to devise a solution to these (all at service class zero) I wonder why Tucker’s capex used is brownfields. They also include FW premise which will now receive FTTN.

        Opex increases under the new model have three components; opex brought forward by faster rollout, additional backend expenses due to more complicated network, higher ongoing costs of active technology alternatives.

        Faster rollout considered a good thing. Increase cost offset by earlier revenues (though I suspect it’ll never reach breakeven).

        Any systems change (eg IT) back to a FTTH network will already have been incurred (ie unrecoverable), leaving the difference in pillar to premises copper line maintence and additional power. We can put figures to both.

        Experience with line maintenance cost varies considerably (AT&T said little, Chorus and Verizon say 50% saving). Exchange to premises copper line maintenance is around $100 per line per year. More than half of that copper is removed in a FTTN network, being generous we’ll use $75. Being generous again a 67% saving is $25 (FTTH) vs $75 (FTTN) per year.

        A 7330 node uses about 1kw/hr to service some 200 premises. At residential rates this is several dollars per year.

        Therefore opex savings (why didn’t Tucker give figures?) of $50 per year dwarfed by the upfront capex differential of $1100. Even assuming a 0% discount rate you’re talking 22 yrs to breakeven. At the industry discount rate of 10% payback is indefinite.

        I assume the higher revenue Tucker identifies is the 30-odd 1gbps of NBNCo’s 1.3m paying customers. 77% of NBNCo’s customers are on 25mbps or below. Again any small increase in revenue dwarfed by the addition cost to service the higher capex.

        Last mile peak speed is actual a small contributor to NBNCo’s revenue model. CVC is more important. NBNCo should require RSP’s publish their contention ratios; much of the “higher speed” revenue would then be captured.

        Contrary to Tucker’s claim FTTH rollout pace has increased massively since the election held in 2013 (doubling premises passed every 12 mths). Few (Tucker and ?) believe FTTH could be completed by 2020-21 (not even Quigley). NBNCo and their construction partners have experienced significant skills shortages and diffuculty ramping up. Where does Tucker propose all the fibre fairies will come from.

        Argue the additional cost is worth it but to claim costs or time are near equivalent is demonstrably false.

        • Richard really isn’t that bright is he.
          I can’t help but pity the intellectually incompetent at times like this.

        • “opex brought forward by faster rollout”
          That faster rollout that has seen existing FTTP rollout slow to a crawl/complete halt now and all other portions of the MTMess suffer a 1.5 year delay on top of the delay that was already expected?

          “I assume the higher revenue Tucker identifies is the 30-odd 1gbps of NBNCo’s 1.3m paying customers.”
          I would assume him to mean any customer on something higher than a 25/10 plan, since anything higher than that is what FTTN will fail to deliver to more than 10% of customers.

          “bean coutbers typically good at both ‘B’ and ‘A’.”
          Rest of post completely C and devoid of B.

          • @Hotcakes: “That faster rollout that has seen existing FTTP rollout slow to a crawl/complete halt now and all other portions of the MTMess suffer a 1.5 year delay on top of the delay that was already expected?”

            As I mentioned in the thread on TheConversation where this article was first posted …
            * In the first four weeks that NBN started reporting weekly rollout data (8 Dec 13 – 31 Dec 13) they made 30,881 services active.
            * In the most recent four weeks (to 19 Nov 2015) they made 83,519 services active.

            The figures for w/e 26 Nov are out now and they hit 90,212 for the four weeks to that date. Their run rate is now 300% what it was when the board was replaced and that is without lying about progress by padding out numbers with SC0.

      • So Richard with the OPEX saving as With you figures is FTTN $340m vs $120M for FTTP.

        5 years down the track to FTTN $1.7B vs FTTP $600M that’s a $1.1B saving in 5 years.

        Delivering faster speeds you you can not give an eg of any ISP offering a min 25Mbps or even a min 50Mbps

  3. Had they not messed with the original NBN plans in 2013 and sat on their hands for 2 years, we would be on our way to being mostly FTTP by the end of the decade. These Liberals have wasted large amounts of time and money for very little gain. Thanks for nothing.

  4. This just highlights that the difference is so small (yes LNP members we the “fibre fanboi’s” admit there is a difference) that its worth spending the little extra to get such a superior technology sooner (that’s sooner as in we’re going to be upgrading to fibre after anyway).

    Yes ‘billions’ sounds large and scary. Over 20 years however its a very small % of this countries GDP and given its an investment and not a cost I’d say its worth it!

    • We spend ~40 Billion per year on our military so 29 Billion Government equity invested in the NBN over 10 years is chump change!

      • Even more so since we’d be ‘investing’ and not ‘spending’.
        eg we get our funds back + interest etc.

        if it were to cost double then sure better to take a good long hard look if it was worthwhile or not with an interim step but its not going to cost double its looking like an extra 10-20% investment.

      • @do is the NBN going to defend Australia? What if a similar level of defence was possible with less money?

        • I don’t have a problem with us spending 40 billion per year on defence and I have even spent time in the defense industry.

          The fact remains tho, military forces don’t generate revenue to pay for themselves or provide a platform for business big and small to boost their productivity, the NBN would have had it not been “interfered with”by the liberals!

  5. Great article (bookmarked for future reference) many points what we predicted before the coalition clowns took over and made it a complete mess. FttP end goal behind schedule and future governments can only make the best of a bad situation due to their delay tactics and breathtaking incompetence.

    • inb4 Richard comes in to claim “high speed internet is the goal, not fibre” without saying what high speed actually is.

      • Yet I had Richard admit that NBN will have trouble delivering any high speeds on FTTN as he couldn’t find an ISP that delivers a min 25Mbps.

        So he contradict his own argument

  6. Not so sure about the extra revenue actually. I’m in a regional area and around 6 months ago one of the ‘retail’ precincts had FTTN finally completed & switched on. To my knowledge at this date, not one shop there has actually signed up to it & in fact had to negotiate not to lose old lines. It is all small business, mostly one or 2 person type shops and they cannot afford the over $5000 they are being asked to actually connect? Some, like a friend of mine who is in a very old building, would have to negotiate with Landlord to even get new lines in the old building and as the owner is down south and could not care less that is not going to happen. So she is still working on her old copper line, which means when we have rain, she loses the connection to her eftpos and has to ask clients to walk down the road to go to the ATM for cash to pay for her services? This is 2015 and an area NBN co has ticked off as done & dusted. Wonder if there are more like this?

    There are a lot of discussions about FTTP vs FTTN etc, but I wonder what the reality is down on the ground? ie How many homes/business have actually been connected, are working, could afford to connect and I mean since the Liberals came into power?

    As the scenario I gave you above changed drastically once the Liberals came in? When the spokesman chatted to the Chamber of Commerce it was discussed how would be similar to current (at the time) Telstra Business broadband costs for business to access (around $125-150) per month maximum, no massive upfront connection cost. Liberals got in, delay happened and type of NBN changed after the delay and now that it is turned on, we don’t know the per month use cost as no-one can afford the thousands to sign up to it?

    Would be great to see a comparison of on the ground costs for the various options for both business & home owners both pre & post Government versions?

    • This sounds weird.

      FTTN uses the old copper, so if you had that deployed the cost to connect would be the cost of a vdsl modem, (usually included in a small upfront cost or subsidised over the life of a multi year contract).

      If you had FTTP deployed, they would have installed an NTU as part of the government controlled installation, now perhaps this NTU under the landlords control and you are talking about needing to run Ethernet to this location?

      Even that shouldn’t cost 5000 dollars. I know we are cbd here but the sparkies wouldn’t charge over 1000 dollars for running Ethernet in the walls of even our most difficult to cable building at work. So 5000 sounds like a fibre pull cost.

      Your example sounds like a scenario that has very little to do with the NBN or is informed entirely by articles in The Australian that quoted prices like 5000 dollars to connect to the NBN back in 2012 (which actually represented an overinflated quote to Ethernet up an entire house).

Comments are closed.