After years of dithering over the value of its Clearwire investment, and undervaluing that company’s treasure trove of 2.5GHz spectrum, Sprint has finally been given a kick in the pants by its new majority owner, Softbank. The US carrier took full control of Clearwire earlier this year, a deal widely assumed to be a condition of Softbank’s investment, and now it has unveiled detailed plans to harness the high capacity frequencies for its ultra-fast LTE service, Spark.
Softbank, unlike Sprint, has always recognized the value of combining high capacity, unpaired spectrum, to boost the speeds and data capacity while achieving coverage in lower frequencies (1.9GHz and 800MHz for Sprint). The Japanese cellco has its own TD-LTE investment at home, through its acquisition of Willcom, as does rival KDDI, via its UQ joint venture. Now Sprint is exploiting the same advantage. Its unusual combination of frequencies will certainly raise issues of device availability, roaming and aggregation methods, but in return it has access to amounts of spectrum which are sufficient for it to make a real impact on the US 4G scene.
This will help it differentiate itself from Verizon and AT&T, which are well ahead in their LTE roll-outs. Sprint currently relies mainly on retaining unlimited data plans ‘ which the other major operators have dumped to save the strain on their networks ‘ and Clearwire’s raw capacity will help it to continue to do this. It will also be able to leapfrog the big two in headline peak speeds, which will grab headlines even if real world performance remains to be tested.
The new Spark service will start at a peak of 50Mbps to 60Mbps and will evolve towards 1Gbps, which would make it one of the world’s first mobile services to qualify to be ‘true 4G’ under ITU definitions. The operators says its spectrum could deliver a theoretical peak of 2Gbps in future. This is achieved by using more and more of the 120Mbps in unpaired 2.5GHz which Clearwire owns in many markets ‘ by contrast, the other Sprint LTE networks use 10MHz and have a peak of about 37Mbps. Not all the 100MHz-plus is being harnessed ‘ some is fragmented and some is still supporting the legacy WiMAX operations, but enough will be active to offer some superfast offerings.
Spark has initially been launched in five markets – New York, Los Angeles, Chicago and two Florida cities, Miami and Tampa. The offering makes its debut with a select group of handsets, such as the G2 from LG and the One Max from HTC ‘ two vendors which will be eager to ride on the profile generated by these announcements. The ubiquitous Samsung is there too, with the Galaxy S4 Mini.
Although the details of Spark, such as pricing, are sketchy so far, it shows Sprint grabbing the PR initiative at last, and it is the first tangible result of some important strategic decisions of recent years ‘ to invest in the flexible Network Vision infrastructure, which supports multiple air interfaces and spectrum bands; to take full control of Clearwire and fully deploy its spectrum rather than sidelining the operator; and to switch off iDEN, while harnessing three bands simultaneously, despite the device complexities, for 4G.
Network Vision will give Sprint the flexibility in future to host other companies’ networks ‘ as it intended to do for now-bankrupt LightSquared and could still do for Dish ‘ and to introduce yet more spectrum bands to its eccentric mix. CEO Dan Hesse indicated at the press conference that Sprint would look to use partners’ spectrum in future. To build out its complex and highly strategic architecture, Sprint is showing some of the habits for which it was once famous, of experimenting with all kinds of technologies and spectrum in order to gain a performance advantage ‘ years ago its Nextel half was the only major carrier to adopt Motorola’s iDEN and, later, to trial WiMAX and other early mobile broadband systems. Now Sprint is continuing that tradition as it battles to stem its still-ongoing tide of customer defections and to narrow the gap with the big two.
That has led it to take an approach more often seen in Asia ‘ and no doubt encouraged by Softbank ‘ of forming a tight ecosystem around itself and partnering closely with its vendors on developments that may sometimes be semi-proprietary or at least heavily customized. That has an impact on cost which must be justified by the performance returns. The three original suppliers to Network Vision were Ericsson, Sprint’s largest CDMA supplier via its acquisition of parts of Nortel; plus Alcatel-Lucent and Samsung, also incumbent vendors via CDMA and WiMAX (Clearwire) respectively. However, Nokia NSN announced this week that it had also got a key role in the Sprint platform, supplying TD-LTE infrastructure alongside Samsung and ALU. This is a key breakthrough for NSN, which has lagged well behind Ericsson and ALU in LTE wins in the US. It was an early supplier to Sprint’s WiMAX roll-outs but exited that project and appeared to be out of the picture at the third US cellco ‘ though it did score LTE wins at T-Mobile USA. NSN has been talking up its strengths in TDD systems, partly the result of acquiring Motorola’s mobile networks business, and this deal confirms that it has moved well ahead of arch-rival Ericsson in this area ‘ which will be significant as carriers increasingly turn to TDD spectrum to support high capacity 4G.
Sprint plans to bring the Spark upgrade to 100 of the largest US cities over the next three years, clearly focusing its Clearwire spectrum on what it does best, delivering capacity in dense urban areas. It also aims to cover 200m POPs with regular LTE by the end of this year and 250m by the end of 2014. It now has 26,000 Network Vision base stations in action, up from 20,000 three months earlier.
Sprint will complete the modernization of its 1.9 GHz network by mid-2014, according to the company, while 800MHz deployment has begun on the spectrum freed up by the shutdown of Nextel’s legacy iDEN services. Most of that low frequency roll-out will take place in 2014, while Sprint will have 5,000 2.5GHz TD-LTE sites working by the end of this year, in line with Clearwire’s build-out plan before it was taken over.
“We continue to believe Sprint can take significant share once they have deployed their 2.5GHz spectrum giving them a speed and capacity advantage; however, it could take longer than expected,” New Street Research analyst Jonathan Chaplin wrote in a research note.
Given that there are relatively few applications which really need gigabit speeds on small screens, the real world fight for the superfast user experience is still wide open. Verizon Wireless is about to bring an additional 40MHz of spectrum, in the AWS band, into its 700MHz coverage network, to boost urban speeds to a possible 100Mbps or more, while AT&T has plans to use AWS and carrier aggregation next year too. All three will be applying some of the techniques which are standardized in LTE-Advanced, including interband aggregation and advanced MIMO antenna arrays, to enhance speed and capacity further.
Of course, eye-catching peak speeds, and the huge capex investment that enables them, can only be justified if the users respond. Sprint is still engaged in its long struggle to stop customers leaving, and having turned off iDEN, it will not be able to blame the defections on declining interest in that legacy service for much longer. In the third quarter, the operator reported a net loss of 313,000 customers, and worse, the loss on the high value, contract side was 360,000. This suggests that, although Sprint has been throwing unlimited data and expensive LTE at the postpaid base, its growth is still coming from the lower value prepaid brands such as Boost and Virgin Mobile.
Hesse warned that the losses would continue into the fourth quarter as it continues to feel the effects of the iDEN shutdown, so the real test will come in the early months of 2014, when that factor should largely have gone away. He also said Sprint was experiencing higher Sprint platform churn because of the Network Vision network modernization project, which has affected CDMA call quality and coverage ‘ worryingly, this is expected to continue well into next year, giving the cellco a new challenge of winning back dissatisfied users. Total retail postpaid churn was flat at 2.09% while retail prepaid churn was slightly up year-on-year at 3.78%. Sprint platform postpaid churn was 1.99%, up from 1.88%.
Despite the customer disappointments, Sprint did turn in its first quarterly net profit since 2007 in the quarter, reversing last year’s Q3 loss of $767m with a gain of $383m. It sold 5m smartphones during the quarter, including 1.4m iPhones, down slightly from 1.5m a year earlier. Smartphones now represent 92% of its postpaid base, its highest percentage ever and that helped deliver Sprint’s best-ever postpaid ARPU at $64.28, up from $63.21 in the year-ago period. Sprint platform prepaid ARPU was $25.33, down from $26.19.
Overall net operating revenue fell 2% year-on-year to $8.68bn and included the best ever Sprint platform service revenues of $5.83bn, up from $5.62bn, while wireless service revenue grew for the thirteenth quarter in a row to hit $7.3bn. EBITDA was up 5% to $1.34bn.