Broadband wireless gained a bad name in the US in the late 1990s when companies such as Teligent failed spectacularly. But WiMAX is ushering in a brand new economic picture for such operators. Not only will standardization push down prices by fostering competition and volume, but spectrum suited to the technology is being auctioned cheaply in many countries in a bid to stimulate broadband - plus there is an unlicensed option. Another key factor is the falling cost of fiber backhaul, with dark fiber now within reach of midrange service providers as well as some enterprises. All this combines to put huge pressure on the financial assumptions of the 3G community, especially in Europe.

It would cost $3bn in equipment, tower, sites, labor and set-up costs to build a national US WiMAX network reaching over 90 per cent of the mainland population. This is an estimate, and a credible one, from In-Stat, and indicates how far the promise of standards, and new spectrum and backhaul options, are shifting the economics of the broadband wireless last mile.

Last time we saw major plans to build out broadband wireless on a national scale was at the turn of the century, with ill-fated ventures by the likes of Teligent, NextLink and Winstar. Comparison of likely WiMAX projects with their cost structures and business models indicates how far the industry has developed, making it a real possibility to deploy wide ranging wireless networks for a fraction of the cost of third generation cellular.

Equipment costs

Three factors are key – the promise of lower cost subscriber equipment and base stations, driven by the greater competition and commoditization that standards engender; new spectrum options in many areas, especially Europe; and the falling cost of backhaul. None of these factors was in the operator’s favor in 1999, nor was the demand for broadband well established six years ago.

The impact on equipment prices of WiMAX will really be felt from next year, once gear is certified and starts to bed down in the market. But in anticipation of the greater competition, and the bigger opportunities, broadband wireless manufacturers are already driving costs down, as Motorola’s sub-$300 pre-WiMAX Canopy subscriber equipment illustrates. The average CPE prices for LMDS gear were $4,000 a unit and a base station was $100,000, while even pre-WiMAX gear is coming down to $500 per CPE and $15,000 per base station. Teligent spent $1.3bn in a year building its initial network in selected cities but only signed up 35,500 customers by the time it filed for bankruptcy in 2001.

Line of sight was a killer for broadband wireless because of the requirement for expensive and disruptive truck roll and the difficulty of creating effective coverage in urban areas. Another reasons why costs outran revenue potential was the lack of standards. Proprietary equipment costs were high because kit had to be made separately for each operator and so vendors never got the scale to reduce prices and so take on wireline infrastructure, especially in an immature market. Equipment makers had to create complete end-to-end solutions, so there was no potential for commodity subscriber units, nor would LOS technologies permit portability, limiting the scope of the applications and their usefulness to businesses.

Spectrum costs

In spectrum terms, WiMAX operators have the option of unlicensed 5GHz bands, though these bring some quality of service risks and make it impossible to exploit the mobile and non-line of sight potential of 802.16 in lower frequencies. A far more attractive choice is 3.5GHz (or 2.5GHz MMDS in North America), although the potential opening up of further, lower bands for broadband wireless could create still greater opportunities in the coming years. For US operators, the shortage of MMDS is a problem – although its once restricted usage is almost certain to be liberalized by the FCC, nearly all the spectrum is held by Sprint-Nextel, BellSouth and Clearwire. All three of these, however, have the financial and marketing resource to build a national network and hardly notice the $3bn bill, since the spectrum is already paid for (and was a low cost purchase to start with, unlike the LMDS licenses that Teligent and the others acquired).

In Europe, many countries are auctioning 3.5GHz licenses this year or did so in 2004 and, as regulators look to extend broadband access to most of the population in line with EC and national guidelines, the trend is to offer the licenses at relatively low cost. For instance, the national licenses for Austria cost about €700,000, compared to the €5m many bidders had originally built into their business plans. In other countries, such as Ireland and some of the new EU entrants, regional licenses can be had for five-figure sums. This has opened up the market to new operators, such as Altitude in France and WiMAX Telecom in Austria, which do not need to raise vast sums in order to become national telcos.

NextLink spent $695m purchasing LMDS licenses from WNP in 1999 and LMDS license costs are estimated to have been $40 per head of population at that time. By contrast, for WiMAX, unlicensed spectrum is free and 3.5GHz is going for under $5 a head, and sometimes under $1.

Backhaul and fiber costs

The third element is backhaul for the networks. One advantage for established telcos seeking to deliver broadband wireless is, of course, their ownership of fiber. In the US, Sprint-Nextel, even if it sells off its long lines business (and if it is not acquired by Verizon), is sure to retain access to its networks on a preferential basis. For Clearwire, the acquisition of AT&T by SBC is a setback – the start-up was almost certain to be the partner in AT&T’s planned national WiMAX roll-out, leasing its emerging network and spectrum in return for access to AT&T’s advanced fiber installations for backhaul.

Even for new operators, however, the backhaul costs are far less daunting than they were in the days of Teligent. Costs of fiber backhaul have collapsed since the telecoms boom and bust. This was because of tightened telco budgets for build-out, increasing competition from wireless backhaul alternatives and lower overall demand. The cost of owning one’s own fiber is now viable for larger companies – once unthinkable for anyone but a large telco – and this is putting further pressure on the prices for leasing operator fiber.

Contrasting, again, with the unsustainable business models of the earlier broadband wireless operators in the US, we saw Winstar involved in huge deals to acquire or lease dark (unlit) fiber to backhaul its network. Winstar spent $640m for 15,000 route miles of dark fiber in a seven-year leasing deal, one of the largest expenses of its roll-out.

The situation is very different now. Indeed, we see some metro area service providers and large enterprises choosing to buy their own dark fiber rather than lease it, as this becomes more cost effective. Steve Corbato, director of network initiatives at the US Internet2 project, commented: “The telecom boom of the late 1990s led to a glut in fiber assets, and the subsequent bust put undeveloped fiber on the market at bargain basement prices.” Gannett, a newspaper publisher that has built its own metro area fiber systems in three cities, said: “The fiber and the equipment are so cheap now, and anyone who is familiar with IP networking gear can handle a short distance optical network.”

Ford is one of the highest profile enterprises to acquire its own dark fiber to support an ambitious converged network strategy. The auto maker claims that owning fiber gives it the bandwidth it needs to support advanced all-IP applications and saves at least 30 per cent on leasing fiber. For a wireless operator, the potential savings could be even greater.

One telco had been leasing five DS3 (44.736Mbps) circuits for voice and data traffic at a cost of $12,500 per month, but has now signed a long term lease on dark fiber for $7,000 a month and built its own backhaul network, with one-off equipment costs of $50,000, that delivers 1Gbps.

Pressure on 3G economics

All this is shifting the economics of broadband wireless rapidly and putting even more pressure on the ROI potential of 3G. Although mobile WiMAX networks will incur additional costs – unlicensed spectrum will not be an option, and the increased volume of traffic will require more base stations and backhaul – the average revenue per user is expected to rise significantly with mobility because the high bandwidths will be able to deliver basic services at low cost to the operator, and to support innovative new applications that will command a price premium among enterprises and power consumers.

This contrasts with the likely pattern for 3G, whose bandwidth ceiling will exclude its operators from some of the most financially attractive services, forcing them to rely on core applications, including voice, that are steadily losing margin. Enhancements like the HSDPA upgrade for W-CDMA, and the EV-DO and EV-DV iterations of CDMA, will help but will require additional investment on a greater scale than building an entire broadband wireless infrastructure.

Like the first wave of national broadband wireless providers, the 3G operators, particularly in Europe, have put massive upfront investment into their systems and gambled on the principle that ‘if we build, they will come’. However, that will now be tested by the challenge from alternative networks that support not just phone-based media applications such as video chat, but the full mobile triple play – a far more powerful attraction than phoneonly services to consumers, in the three-year timeframe required to get the subscriber units right, and allowing the operator to deliver significantly more bandwidth for relatively low cost because of moderate set-up and running costs and the flexibility of bundling packages.

Researchers forecast that by 2006, there will be 20.6m active Wi- Fi users and 831,000 WiMAX subscribers in Europe versus 21.2m 3G subscribers.

Many of these, of course, will have at least two of these subscriptions but, though the average spend on wireless data services will rise steadily, a multi-network subscriber will spend less on 3G than a 3G-only user (in other words, WiMAX will not be purely an additional spend). Indeed, if the mobile operators persist in billing out 3G services by the megabyte instead of a flat fee model, more people will go for Wi-Fi – yet as they make the inevitable shift to all-you-can-eat, margins will suffer.

On data, broadband customers are traditionally used to spending $30 to $60 per month, more in rural areas or for premium services like video on demand. While this figure will erode, the mobility aspect of WiMAX should support a similar rate in 2007, plus high value revenues from enterprises. This could impact the expected data revenue from 3G, which is expected to be about €16 a month by the end of 2005 in western Europe.

Although there is currently strong association between adopters of broadband and of 3G, the ability of WiMAX to span both camps could make 3G data less attractive or drive its operators to low, flat pricing to compete. Yet Analysys estimates that the monthly ARPU required for a European 3G operator to gain ROI on its network investment will be, in 2005, about $25 from voice and $30 from data, with the voice figure declining to $20 by 2010 but data rising to $60. This will be extremely hard to achieve with the impact of flat rate pricing and VoIP, and a new estimate from Telecom View says that, by 2010, monthly ARPU for 3G in both voice and data will be around $50.

Copyright © 2004, Wireless Watch

Wireless Watch is published by Rethink Research, a London-based IT publishing and consulting firm. This weekly newsletter delivers in-depth analysis and market research of mobile and wireless for business. Subscription details are here.

Related stories

3G success hangs on handsets
WiMAX delay shakes investor confidence
Make-or-break year for WiMAX