After reading the NBN business plan it’s actually a bit hard to remain aloof and unexcited. This is a magnificent, awe-inspiring undertaking: there has never been anything like it, not in this country and probably not anywhere in the world. The scale of the project is staggering. They are going to lay fibre optic cable past 5000 homes and businesses every day, day after day, for years on end, in the most sparsely populated land on Earth and end up connecting every premise to either fibre, wireless or satellite. It’s a total rejig of a nation’s communications infrastructure, top to bottom. The only negative, in my view, is the one identified inBusiness Spectator by Simon Hackett this morning – that the ACCC’s demand for 200 points of interconnect (negotiated down to 120) will make life very difficult for small operators (NBNCo's point of disconnect, December 21). But getting the number of POIs was always going be a difficult juggling act, and was probably never going to make everybody happy What’s more the revenue plan looks absurdly conservative. It seems to be based on the idea that the growth rate of broadband internet traffic is about to collapse, when in fact it is just beginning to take off thanks to the development of online video on demand and IPTV Not only will the NBN not be a white elephant it will almost certainly prove to be a great investment. In fact, without wishing to get carried away (too late do you think?) it could represent, on its own, a huge national savings plan. When it’s finished the asset will be worth several times the government’s investment of $27.5 billion. As with all business plans, the one released by the NBN Co yesterday is a definite proposal about costs and a pure guess about revenue. The idea with these things is to make the lowest revenue projections you can get away with to justify the capital and operating costs that you want to make. Always best to under-promise and over-deliver. As a result most business plans are reverse engineered. The government told the NBN that it had to produce an internal rate of return (IRR) of 7 per cent to generate a thin but politically acceptable positive return over its cost of funds. Hey presto! The projected IRR for the NBN is 7.04 per cent. What a coincidence! Is that really what it will be? Goodness knows. The exact growth rates of internet traffic assumed in the business plan are not revealed, although there are some graphs showing spectacular forecasts from Alcatel Lucent and Cisco Systems, with the dotted lines of the NBN Co forecasts under them at a much flatter gradient. Looking at the summary financials on page 134, we can see that the number of premises connected by 2020 is assumed to be 7,845,000, from which the NBN will earn revenue of $5.76 billion. On my calculations, that’s an absurdly low $52 per user per month in 10 years time in wholesale revenue from providing all fixed communications services – phone, TV, internet. Doing the same calculation for each year up to 2020 gives an average annual growth rate in revenue per user of 10 per cent. In fact average data usage in Australia grew 30.8 per cent over the past 12 months, according to the ABS, and that’s without video on demand. Cisco is predicting a compound annual growth rate in internet data traffic of 34 per cent per year to 2014. In Hong Kong it has been growing at 68 per cent compound for years. Data growth rates don’t turn into the same revenue growth rates, since prices tend to fall and data plans are capped. Even so, NBN Co’s forecast of 10 per cent per annum growth in revenue per user looks very conservative and so do the traffic forecasts in its graphs. What’s more, all this is only from "addressable premises", described in Attachment A to the government’s "Statement of Expectations" as: locations used for residential, business, government, health or educational purposes. There’s also a long list of  "non-premises" to which the NBN is "permitted but not required" to connect, such as: traffic lights, bus stops, train stations, ATM terminals, traffic and security cameras, mobile phone towers, street lights, water, gas and electricity infrastructure, weather monitoring devices and security systems. Many of these will be very profitable businesses, but they are not included in the revenue projections at all. The growth in internet television that is now taking off, plus profits from all those "non-premises", mean the IRR of the NBN will certainly be a lot more than 7.04 per cent, and the end value a lot more than $25.7 billion. It could be sold in 20 years to create a big sovereign wealth fund. Alternatively we could hang on to it and collect the revenue. Since privatising state-owned telecommunications monopolies has not proved a great idea in the past: you just have to go and build another one to break up the one you sold. *This article was originally published on Business Spectator