US Wireless Carriers: The Barbarians are at the Gate
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Artur Pylak
FOR IMPORTANT DISCLOSURES 203.901.1633 /.1634
sagawa@ / firstname.lastname@example.org
April 2, 2013
US Wireless Carriers: The Barbarians are at the Gate
- The US wireless industry is a thirty year story of haves and have-nots. Spectrum holdings, access to capital, and strategic misfires have yielded an oligopoly with VZ and T exploiting inherent cost and coverage advantages to perpetuate high prices, low capex, restrictive usage policies, and supernormal returns. However, M&A, new spectrum, technology, and deep pocketed 3rd parties threaten the status quo, even as user priorities shift from blanket coverage to fast data. Long time sad sack Sprint has cash from Softbank and a plan for Clearwire’s spectrum. T-Mobile and MetroPCS get operating synergies and enough spectrum for a competitive 4G network. New technologies – e.g. LTE Advanced, small cells, multi-band radios, etc. – will also help level things by allowing carriers to boost capacity, increase speeds, and lower costs. The secondary carriers are deploying 4G in new bands, and the market awaits a new swath of 700MHz to be auctioned in 2014, under FCC rules which could stymie VZ and T. These auctions could also bring in 3rd parties, like Google, with strong interest in fast, cheap, limit free wireless. Meanwhile, the spectrum-based coverage advantage that has separated haves and have-nots is growing less important to users who crave fast and reliable data. Together, we expect these factors to amp up competition, leading to more network investment, lower prices, and fewer usage limitations, not only to the benefit of mobile users, but eventually playing for residential broadband as well.
- VZ and T have dominated US wireless. The two 850MHz licenses, granted in 1981 and consolidated into VZ and T through years of M&A, carry substantial advantages vs. the higher frequencies that have been subsequently issued to other carriers, enabling better coverage from fewer cells, and thus lower costs. In addition, synergy and cash flow from wireline operations fueled more aggressive build outs than poorly capitalized rivals. Finally, smaller carriers made some serious strategic errors – e.g. Sprint’s pricey Nextel purchase and its commitment to the dead-end WiMAX technology. These advantages have allowed VZ and T to maintain unusually high prices, while damping usage with tiers, caps and throttling, generating oligopoly rents.
- Softbank is rescuing Sprint. After watching Nextel die, and giving up on WiMAX in favor of LTE, Sprint was overleveraged and facing pressure to divest its stake in Clearwire and its large swath of unusual 2.5MHz band spectrum. Sprint was on bankruptcy watch until white knight Softbank stepped up with a $20B offer for 70% of the company. The deal will infuse Sprint with the capital to take control of Clearwire and build out LTE Advanced. Softbank, which holds similarly unusual spectrum in Japan, gets scale for the technology needed to exploit the 2.5MHz band, and a crack at using the aggressive tactics that worked at home in the US market.
- T-Mobile USA and MetroPCS are a great match. Meanwhile, 4th place T-Mobile lacked the scale, spectrum and capital commitment from Deutsche Telkom needed to make headway. After a deal to sell it to T was waved off by the FCC and DoJ, DT turned around and agreed to merge T-Mobile with super-regional player MetroPCS. The combined company would hold 72MHz of spectrum in most major markets, within 2 commonly used spectrum bands, giving ample room to build out LTE with unfettered choice of equipment and devices.
- Technology advances will help level the playing field. LTE updates will increase speeds and cell site capacity, while enabling carriers to aggregate disparate pieces of spectrum and to use “unpaired” spectrum, such as Clearwire’s 2.5GHz. “Small cells” are extremely cost effective for alleviating hot spot congestion and reaching into residential neighborhoods without towers. The newest specs allow small cells to piggy back on macrocells for backhaul, further cutting network costs. Finally, Qualcomm’s innovative 40-band RF chip portends a future where most devices would work on most networks, increasing the availability and lowering the cost for carriers.
- New spectrum will likely benefit the smaller carriers. Sprint is refarming its 800MHz Nextel spectrum to LTE and moving on Clearwire and its huge 2.5GHz band holdings. T-Mobile got spectrum from AT&T as a break-up fee, and will have an average of 72MHz in the top markets after the MetroPCS deal closes, including 40MHz of open spectrum for LTE. Dish Networks is sitting on an additional 40MHz that could be a boon to smaller carriers. Finally, the FCC is moving to auction as much as 120MHz of highly desirable 700MHz band spectrum, likely before the end of 2014 and under rules greatly favorable to smaller carriers and new entrants.
- Deep pocketed internet players could enter. High prices and data caps are anathema to cloud players like Google or Amazon. Both have been rumored in talks with Sprint, and special arrangements with carriers willing to commit to low priced, unlimited wireless data make sense in both directions. With the upcoming 700MHz auctions the likely last chance at a national footprint in prime spectrum, we would be surprised if the Internet players were not involved.
- Consumers care more about data speeds than voice coverage. The first 30 years of wireless competition were about blanket coverage, to the advantage of the incumbent leaders, but the smartphone/tablet revolution brings even greater priority to fast, available data. With LTE Advanced, Sprint and T-Mobile will be able to offer relatively uncongested and unlimited wireless internet into populated areas, without the expense of blanketing the country.
- Residential broadband IS in play. The capabilities of LTE Advanced – 100Mbps user speeds, 2Gbps/cell capacity, $35K base station price, etc. – would allow wireless to compete with wired residential broadband priced at $50+/month. With 70% of American households having no other option but a cable modem for video-capable internet, and the self-interest of cable MSOs to thwart on-line alternatives to their television service, the opportunity to inject competition into this market would be irresistible for internet players, should they enter as expected.
- Legacy carriers will suffer, upstarts will gain, and consumers will win. VZ, T and cable operators will have to cope with newly vital rivals, competing with more comparable cost structures, improving coverage and potentially superior performance. Both Sprint and T-Mobile stand to pick up market share in mobile, and may enter fixed residential broadband. Google, Amazon, Microsoft, Apple, Facebook, and others will see mobile usage rise with falling prices and bandwidth caps. Most of all, users will see lower monthly bills and uncapped usage.
The Revenge of the Nerds
US wireless has been a bully’s game for a long time. VZ and T trace their heritage to the original 25MHz cellular licenses, granted without fee in 1981 in the prime 850MHz band. In 1994, the FCC auctioned 7 chunks totaling 130MHz in the less attractive 1850MHz PCS band, broken into 493 separate geographic areas. Some of this spectrum was acquired by the original license holders to add capacity, but the biggest buyers were new entrants to the market, amongst them, Sprint and VoiceStream, which eventually became T-Mobile after a long string of M&A transactions.
The original license holders had considerable advantages vs. the PCS entrants – e.g. scale economies, the ability to bundle wireless with wireline, superior spectrum holdings, and deep pockets to bid for more. Spectrum was critical – radio waves travel far further and penetrate walls far better at 850MHz than they do at 1850MHz, a fact of physics that allowed the proto-VZ and T networks to deliver better coverage with far fewer base stations. Once the dust had settled on industry M&A, VZ and T had superior national coverage and lower costs than any of the wannabe rivals, allowing them to demand price premiums and aggressively limit their customers’ data usage.
However, this cash flow nirvana of high prices and low capex may be at risk. Consumer priorities may be shifting from valuing blanket voice coverage toward valuing fast speeds and high limits with adequate coverage. T-Mobile’s planned MetroPCS merger brings much needed scale and operating synergies, with enough spectrum headroom to deploy 4G LTE. Sprint’s cash infusion from Softbank will allow them to take full control of Clearwire, just as the technology needed to exploit its massive holdings of spectrum in the unusual 2.5GHz band is becoming commercially available. Indeed, technology advances are a major leveler, making it cheaper for smaller carriers or new entrants to play network catch-up.
Moreover, new spectrum is coming available. As much as 120MHz of the prime 700MHz band will be auctioned before the end of 2014, with the Democratic FCC expected to set bidding rules limiting spectrum-rich VZ and T’s participation. Dish network is sitting on 40MHz adjacent to the AWS band, and, while it has made a play for Clearwire, has been rumored to be in talks with Sprint on some sort of tie-up. We also note that on-line players, particularly Google, but also Amazon, Microsoft, and Apple, have the motivation and the money to invest in promoting competitive wireless broadband with low prices and no limits. That LTE Advanced promises to make wireless networks more than competitive for fixed residential broadband would be more icing on the cake to tempt these outsiders from the sidelines.
The biggest winners from this scenario would be US consumers, who could see lower prices and more generous usage terms for wireless and residential broadband. Internet businesses – Google, Amazon, Facebook, et al. – also stand to benefit. Smaller carriers, like Sprint, T-Mobile, and the regional players, will gain footing to compete, with more room for resellers, such as Virgin Mobile, to prosper. The obvious losers are VZ and T, but residential broadband operators would be affected well before the end of the decade.
The idea of radio telephones in automobiles dates all the way back to 1947 and AT&T’s Mobile Telephone Service. The user equipment weighed 80 pounds and calls were manually connected by operator to vehicles which needed to remain in range of a single tower. The idea of a cell-based network with hand-offs as users moved from cell to cell didn’t emerge until the late ‘60’s, and the possibility of truly portable user devices was not demonstrated until 1973, when Motorola engineer Martin Cooper used a 2.5 pound handheld prototype to call Dr. Joel Engel of Bell Labs from 6th Avenue and 54th street in New York City. Commercial services based on the cellular approach were launched in 1978, prompting the FCC to get serious about setting aside a standard spectrum block for the fledgling technology.
In 1981, the FCC began issuing 20MHz licenses for cellular service in each market (later topped up to 25MHz), giving one to the dominant wireline operator in each area, usually AT&T, and awarding the other based on public comparative hearings. A year later, with the Bell System break-up, AT&T agreed to assign the wireline licenses to the seven newly created Bell Operating Companies, relying on a consultant study which had advised it that the ultimate market for wireless telephones might be less than a million users nationwide. The biggest winner at the comparative hearings was Craig McCaw, who leveraged himself to the hilt to buy licenses that had been awarded to others, adding them to those he was able to win directly. McCaw acquired MCI’s wireless operations in 1986 and LIN Broadcasting in 1989, finally stitching together a roughly national footprint which was branded Cellular One. AT&T bought out McCaw in 1994, and the network was rebranded as AT&T Cellular (Exhibit 1).
Exh 1: Historical Timeline of AT&T Wireless
Meanwhile, the seven Bell Operating Companies (RBOCs) moved forward with their own wireless operations. By the end of the ‘80’s it was apparent that a shift to digital technology was forthcoming, and the U.S. Cellular Telecommunications Industry Association spearheaded the move to establish a common standard format for American operators. The resulting standard, IS-54, based on TDMA technology, was set in 1992, and was intended to counter the GSM standard championed by the European industry on the world stage. However, several US carriers, led by the former Pacific Telesis Wireless (by then split out as AirTouch Cellular) and Bell Atlantic, broke ranks to select CDMA, an upstart technology from the then start-up Qualcomm.
After the launch of digital networks in the mid-‘90’s, the industry was prime for consolidation along the lines of these technology choices. TDMA champion SBC rolled up the independent SNET in 1998, bought Ameritech in 1999, and formed the Cingular Wireless JV with BellSouth in 2000, eventually buying out its JV partner in 2006. Cingular finally bought AT&T Cellular in 2004, and embarked on the arduous task of transitioning its entire network from the TDMA standard to the European GSM, which had established itself as an overwhelmingly dominant global standard with cheaper and better equipment and handsets to show for it. Since SBC’s 2007 purchase of its former parent, AT&T, the wireless operation has been branded as AT&T Wireless.
CDMA champion BellAtlantic bought NYNEX in 1997, and independent GTE in 2000, the same year it formed the Verizon Wireless JV with Vodafone, which contributed the CDMA networks it had gained from its 1999 acquisition of AirTouch Cellular, which had itself swallowed US West’s wireless operations in 1998. Verizon became the largest wireless operator in the country with its 2009 acquisition of Alltel and its 13 million, mostly small market, customers. Because Verizon and its acquisitions had all opted for the CDMA standard, the carrier did not face the same wrenching transition to GSM borne by AT&T, giving it superior service quality and substantial market momentum during the better part of the last decade (Exhibit 2).
Exh 2: Historical Timeline of Verizon
Along the way, as both Verizon and AT&T consolidated their national footprints through their serial acquisitions, both carriers and their eventual merger partners gobbled up additional spectrum as it was offered by the government. In 1994, the FCC began auctioning off new blocks of spectrum to be used for commercial cellular service, starting with the 1850 MHz “Personal Communications System” or PCS spectrum. Initially, PCS was expected to be a limited mobility service – e.g. incomplete coverage, no vehicular speed call handoffs, etc. – with a much lower price point intended to bring portable voice communications to the mass market, but that notion was quickly scrapped and the original cellular license holders became major bidders for the PCS spectrum, even enlisting shell entities to bid on their behalf for licenses set aside in the auction for minority-owned small businesses. AT&T and Verizon were also major bidders at the 2006 auctions for another major block of spectrum in the 2.1GHz band dubbed “Advanced Wireless Services” or AWS, with Verizon picking up another big slice with their 2012 acquisition of the AWS licenses originally purchased by a consortium of cable operators. Finally, the two major incumbents were also the biggest winners at the 2008 auction for spectrum in the exceptionally attractive 700MHz band that had been freed up with the digitization of broadcast television (Exhibit 3).
Exh 3: US Mobile Wireless Frequencies and Auctions
Up High, Down Low, Too Slow …
Sprint and T-Mobile had their genesis in the 1994 PCS auctions. At that time, Sprint was America’s 3rd largest wireline long-distance operator, tracing its roots back to an independent telephone company founded in Abilene Kansas in 1899. Sprint was an aggressive bidder for PCS spectrum, consolidating many market licenses into a near national footprint at 1850MHz. A year later, Sprint acquired independent telco Centel, which had both wireline and wireless operations spread across 18 states. Sprint joined Bell Atlantic, AirTouch and GTE (all later consolidated into Verizon Wireless) in championing the fledgling CDMA technical standard, gaining some momentum during the ‘90’s from its broad national footprint and aggressive marketing. However, with the turn of the millennium, AT&T and Verizon’s march to their “manifest destiny”, combined with the superior coverage enabled by their inherently better 850MHz spectrum, had put Sprint on its heels (Exhibit 4).
Exh 4: Historical Timeline of Sprint Nextel
In 2004, Sprint announced a merger with Nextel, which had built a highly profitable business providing “push to talk” walkie-talkie-like service to enterprise customers via a specialized network operating in small chunks of 800MHz band spectrum that had been long allocated for commercial dispatch applications. The deal was a classic “buy at the peak” blunder, and not long after, both AT&T and Verizon had deployed analogous push-to-talk capabilities developed for their own network technologies, while the newly dubbed SprintNextel struggled to bridge service across its two fundamentally incompatible networks. The handwriting was on the wall, and despite SprintNextel’s trumpeting the superiority of its push-to-talk performance, it began to hemorrhage subscribers (Exhibit 5).
Exh 5: Decline of Sprint Nextel “iDEN” Platform Subscribers
Meanwhile, Sprint also had a line on a huge, and potentially lucrative block of unused spectrum. Way back in the ‘70’s, the FCC had assigned 194MHz of spectrum between 2495MHz and 2690MHz to educational institutions for the purpose of providing instructional television services. This band was underutilized for years, and around the turn of the century, the FCC amended these licenses to allow sub-leasing for two-way commercial communications services. Sprint and the Craig McCaw controlled Clearwire began negotiating leases with the various institutions holding them, and between the two of them aggregated most of the rights to the band. In 2003, the FCC again amended its rules to specifically allow wireless broadband services, and instituted a “use it or lose it” rule that required license holders to have deployed a network to substantially use the frequencies by 2011 or risk having the licenses reclaimed by the FCC.
In 2006, under the gun to begin deploying their networks and hoping to get a jump on Verizon and AT&T, which were planning their own 4G networks, Sprint and Clearwire signed deals with Intel, Motorola and Samsung providing financing for a network based on WiMAX. This technology was pitched as having an 18-24 month lead on the LTE standard championed by most of the telecommunications industry, and Sprint and Clearwire believed that the performance advantage would allow them to leapfrog the market leaders.
Alas, it was not so. WiMAX was late and not all that much faster than the 3G services offered by rivals. In 2008, Sprint agreed to exchange its newly built WiMax network and spectrum rights for a major stake in Clearwire, signing on to distribute the merged service under the Sprint brand. Most semiconductor companies and device makers didn’t invest in the WiMAX standard, and initial service was limited to lap-top cards for years, with the first commercially available dual-mode handset not available until mid-2010, just 6 months ahead of Verizon’s LTE service launch. By 2011, Clearwire threw in the towel and announced that it begin building out its own LTE network, effectively sounding the death knell for WiMAX after 5 years and billions in investment. Clearwire flirted with bankruptcy, and many analysts expected Sprint to follow.
For its part, Sprint also announced in 2011 that it too would begin deploying LTE, in the patchwork frequencies previously used by the now defunct Nextel service. Both of the new LTE networks, Clearwire’s and Sprint’s, relied on a new release of the 4G standard that allowed for “unpaired” spectrum by using time division techniques to use the same channels for both upstream and downstream communications. Finally, Sprint had the right technology, but after the Nextel and WiMAX debacles, deploying the networks would require investment capital that was in short supply.
Was it Over When the Germans Bombed Pearl Harbor?
The T-Mobile USA story is far less dramatic than Sprint’s. The carrier began life in 1994 as VoiceStream, launched by its then parent Western Wireless, as a vehicle for building a digital network in the PCS spectrum auctioned in that year. VoiceStream was spun off to shareholders in 1999, and in 2000 merged with fellow PCS operator OmniPoint, and acquired regional PCS player Aerial, establishing a near national footprint for its GSM based network. In 2001, VoiceStream was acquired by Deutsche Telekom, which renamed the operation T-Mobile USA, in line with its global wireless brand. T-Mobile added a national license at 2.1GHz to its 1850MHz band during the 2006 AWS spectrum auctions, and used it to build out an enhanced version of 3G WCDMA technology that could support download speeds that were comparable to the initial 4G networks launched by its rivals. Despite scale disadvantages, T-Mobile has competed as a bargain carriers dangling low prices and ample data usage to gain subscribers, and leading in the low-end prepaid segment (Exhibit 6).
Exh 6: Historical Timeline of T-Mobile
Can You Hear Me Now?
By 2012, the US wireless market had fully settled into oligopoly mode. AT&T and Verizon each controlled more than 100MHz of spectrum in the majority of US markets, anchored by their holdings in the coveted 700MHz and 850MHz bands. Excluding Sprint’s star-crossed stake in Clearwire, the two smaller national carriers managed with less than 70MHz apiece, most of that at frequencies of 1850MHz or higher (Exhibit 7). The laws of physics dictate that lower frequency radio waves take longer distances to decay and penetrate through solid objects, such as walls and trees, more easily than do higher frequency radio waves. At 850MHz, the practical range of a single macro base station is nearly double that of a cell site operating at 1850MHz, reaching an area as much as 6 times larger (Exhibit 8).
Exh 7: Carrier Spectrum Holdings Summary
Exh 8: Relative Coverage Advantages of Lower Frequency Spectrum versus Higher Frequencies
Exh 9: US Wireless Carrier Metrics, Q4 2012
Essentially, Verizon and AT&T could cover more territory with fewer base stations than either Sprint or T-Mobile. The two leaders could brag of better signal coverage and demand higher service prices, while enjoying scale and signal range based cost advantages (Exhibit 9). Moreover, Verizon and AT&T had nearly unfettered access to capital, while cash poor Sprint and T-Mobile struggled to finance their game of catch-up against their fundamentally advantaged peers. The result? US consumers pay the second highest monthly wireless bills in the developed world while suffering through the indignity of data usage caps, designed to minimize AT&T and Verizon’s capital spending. The top two carriers have 66% of US wireless users, but generate 67% of the revenue and 81% of the profits.
A New Hope
In March 2011, the AT&T death star set its sights on T-Mobile, signing a $39B deal to acquire the GSM carrier, with $6B break-up penalty in cash and spectrum should the acquisition fall through. The proposed deal would have vaulted AT&T well past Verizon into first place with 130 million subscribers and nearly 150MHz of spectrum in most markets. By August, the US DoJ had filed its intention to sue to block the completion of the deal, with the FCC expressing grave reservations that the combination could be competed in any way acceptable to the public interest. By December, AT&T backed off, leaving T-Mobile with the agreed upon penalty as a nest egg for a more aggressive 4G strategy.
10 months later, T-Mobile announced that it would merge operations with the number five US Carrier MetroPCS. While MetroPCS ran its 2G network on the CDMA standard, incompatible with T-Mobile’s GSM 2G network, their 3G assets were compatible and, importantly, their spectrum holdings were highly complementary. Both carriers operate in the PCS and AWS bands, controlling 72MHz on average in major markets (Exhibit 10). Both carriers have begun the difficult process of freeing some spectrum for a 4G LTE build-out, a process that will be dramatically easier as a combined company. The resulting carrier will have better scale economies and enjoy significant operating synergies, yielding more headroom for aggressive pricing, generous data plans, and, importantly, profit.
Exh 10: T-Mobile / MetroPCS Spectrum Holdings
The Sun Also Rises
Prior to T-Mobile’s deal, most industry speculation swirled around a Sprint-MetroPCS tie up, and even after the fact, analysts opined that Sprint could step in with a counter offer, Instead, two weeks later, Japan’s Softbank stepped up with an offer to buy 70% of the ailing Sprint for $20B, with a follow up deal in December for Sprint to take control of Clearwire. This makes sense for Softbank, which has begun deploying a Japanese LTE network in the same 2.5MHz EBS band as Clearwire. Adding a well-financed Clearwire to the Softbank network will make 2.5GHz TDMA LTE far more attractive to equipment and device makers, increasing choice and reducing technology prices in both markets for Softbank. While the EBS spectrum carries a considerable cell size and indoor penetration penalty vs. the lower frequencies employed by AT&T and Verizon, Clearwire holds 154MHz of EBS bandwidth per market on average, more than either of the two leaders in all of their bands combined across all of their technologies and services. This means that subscribers able to get a good signal from Sprint/Clearwire on its planned LTE Advanced network will see far faster speeds with far less congestion than any other carrier in the market. Marketed properly, this could be a big deal.
LTE Advanced … Advances
The wireless industry has a solid record of hitting its technical marks. The 3G Partnership Project or 3GPP took over management of the GSM/WCDMA family of standards from the European Telecommunications Standards Institute (ETSI) in 1999, and has shepherded 3G and now 4G standards to global adoption and steady technical improvement. When first launched in 2001, 3G WCDMA networks barely managed 1 Mbps downlink throughput. A decade later, theoretical speeds topped out at 42Mbps, with 10Mbps downloads typical for well managed networks employing the HSPA+ specs included in 3GPP release 7. Release 8, ratified in 2008, laid out the specs for LTE (Long Term Evolution), a new wireless network standard with a development trajectory that promised to deliver the International Telecommunications Union’s (ITU) definition of 4G service – i.e. 100Mbps mobile download speeds and 1Gbps static download speeds from the same network. The first LTE networks came live in 2010, and delivered peak download speeds of up to 30Mbps in uncongested cells, with average speeds of more than 10Mbps. In the US, Verizon was first to deliver LTE, using a portion of the 700MHz spectrum that it has licensed during the 2008 auction.
3GPP release 10, specifying a major enhancement to the 4G LTE standard dubbed LTE Advanced, was finalized in early 2011, with commercial silicon available this year and network equipment by early 2014. LTE Advanced offers several major advantages over LTE. First, it boosts the speed and network capacity by a factor of 3-5 times, making 100Mbps peak performance realistic (Exhibit 11). Second, it can be implemented in time-division mode, a necessity for carriers with spectrum bands that haven’t been defined with clear upstream and downstream channels – e.g. Sprint/Clearwire. Third, it can be implemented in pieces of spectrum of variable size, from as low as 1.5MHz to as high as 100MHz. Fourth, pieces of spectrum in disparate frequency bands can be managed together as a single network system. Fifth, it further facilitates the use of “small cell” technology – inexpensive lower power base stations designed to fill in coverage holes and provide extra capacity in areas of congestion – and allows small cells to backhaul to the network by connection to a neighboring macro cell.
Exh 11: Wireless Data Standards Releases by Technology
Exh 12: Wireless and Wireline Deployed Network Speeds, 2000-2015
These innovations will make it easier and cheaper for Sprint, T-Mobile and any other would-be competitors to roll out competitive high-speed wireless network capacity, using almost any spectrum that they can get their hands on. This expansion of usable spectrum is also enabled by advances on the device side. Qualcomm recently announced an RF radio transceiver for portable devices that supports 40 different bands in a single chip, covering almost all of the spectrum likely to be assigned to commercial wireless service around the world. This will allow device makers to design and build a single model appropriate for most of the world’s carriers, instead of the 3-4 variants that have traditionally been required. This has obvious advantages for smaller carriers with unusual spectrum holdings, which have historically struggled to find device makers willing to build to their needs.
Please Sir, Can I Have Some More?
More spectrum appears to be on the way. Despite puffed chest posturing by the National Association of Broadcasters, the FCC is proceeding apace to set up incentive auctions for as much as 120MHz of prime 700MHz band spectrum currently licensed to television broadcast stations. Under the proposed terms, station owners would be offered a share of the proceeds of the auction in return for agreeing to give up their current spectrum assignment and shift their broadcasts to a channel that could be shared with other broadcasters. While some broadcasters continue to carp that they have legitimate business opportunities to use the extra spectrum on which they are currently sitting, we expect most will jump at the opportunity to monetize the unused asset. Indeed, the private investors that have been quietly snapping up local television stations around the country for prices that could not be justified by the actual cash flows of the broadcast business see the payoffs clearly.
Exhibit 13: National Broadband Plan Proposed Mobile/Broadband Spectrum
120 MHz is a lot of spectrum – Verizon’s current LTE network runs in 22MHz, with plans to expand into 40MHz of AWS spectrum that it has acquired from a consortium of cable operators. The auction process will likely break the available band into 20-40MHz pieces, with at least some of the licenses offered as a national footprint (Exhibit 13). Recent comments from the FCC suggest a commitment to restricting Verizon and AT&T from adding to their 700MHz holdings beyond filling in geographic holes, although changes at the commission, including a new, yet to be appointed, Chairman, could yield some softening of that policy. Nonetheless, we expect the auction rules, set by an Obama appointed FCC already authorized by Congress, will greatly favor smaller carriers without 700MHz band holdings, like Sprint and T-Mobile, and new entrants. Assuming that the currently envisioned auction timetable holds, the licenses would be assigned by mid-2014 and the spectrum cleared for operation by the end of 2015.
Meanwhile, Dish Network has received approval to use 40MHz of S-Band satellite spectrum adjacent to the AWS band for commercial wireless service. The approval comes with a seven year “use it or lose it” provision requiring Dish’s wireless network to cover at least 70% of the territory by 2019. While Dish could chose to build an independent network with this spectrum, perhaps augmenting it with bandwidth to be acquired in the 700MHz auction, we expect Dish to partner with either an existing carrier, most likely Sprint or T-Mobile, or an interested 3rd party, such as Google or Amazon.
Now Playing Left Field …
High wireless prices and tight usage limitations are a not only a problem for consumers, but also for the cloud-based businesses that hope to serve them (Exhibit 14). As US internet use turns increasingly mobile, the existing oligopolistic carrier market structure is a substantial impediment to the ambitions of companies like Google, Amazon, Apple, Facebook, and Microsoft. Given that these five companies alone have more than $280B in cash and investments on their balance sheets, the financial firepower that could be brought to bear in the cause of cheap and plentiful wireless broadband is awesome.
Exh 14: Minimum Cost of Complete Mobile Service Packages by Country
Increasingly, it appears likely that these players may be contemplating a move from the sidelines. Google has invested roughly $500M in building out a residential fiber optic network in the Kansas City market, offering multi-channel video and 1 Gbps internet service at significant discounts to the incumbent residential players, and introducing a crowd-funding style pre-payment model to help offset its capital requirements. It would not be out of character for Google to make a multi-billion-dollar play for 700MHz spectrum or to forge a deal to help a secondary carrier make a low-price, no-limit play for market leadership. To that end, both Google and Amazon have been rumored to have entered discussions with both Dish and Sprint with regard to participation in LTE Advanced network launches.
I Feel the Need for Speed
For most of the first 30 years of the cellular telephone market, the sole barometer for assessing network quality was “Can I connect a call and be confident that it won’t drop?” Carriers gunned for blanket national coverage and led their advertising with dropped call statistics and coverage maps. On this measuring stick, Verizon and AT&T with their 850MHz licenses were at enormous advantage and it showed in their growing market share lead vs. their intrinsically hobbled rivals. The original license holders could provide coverage along every major highway in the nation far more cheaply and with less likelihood of gaps. Moreover, the 60-70 thousand base stations necessary for blanket national coverage were very expensive, giving well capitalized AT&T and Verizon a further leg up.
Exh 15: Global Monthly Mobile and Voice Traffic
However, by 2009 wireless data traffic had surpassed voice for US carriers, and now accounts for more than 75% of total wireless network usage (Exhibit 15-16). Within that shift is a change in the criteria for measuring the network. Wireless data usage is not sensitive to the temporary signal losses that lead to dropped voice calls, and the traffic is more highly concentrated into “hot spots” subject to congestion. Data users care about their typical download speeds and having ample capacity to avoid vexing slowdowns in busy places at busy times.
Exh 16: Voice Use is Becoming Less Important – Average MOUs per User per Month
The new reality suits the underdogs, who needn’t deploy LTE advanced in 60,000 cell sites to achieve blanket coverage, but can cherry pick extremely congested markets and double down on speed and capacity. The 3G networks already in place can suffice for voice coverage in the hinterlands. In addition, years of competition in the base station market has bought prices down, as has the rise of “small cells”, making the price of a ticket to play lower than it has ever been once you have the spectrum.
Around Third and Heading for Home
To date, the capacity and cost of wireless networks made them impractical competitors to residential broadband, but this will not always be the case. The initial specification for LTE Advanced supports a theoretical maximum aggregated download throughput of 3Gbps per base station, assuming a maximum spectrum assignment of 100MHz, with the realistic capacity in the range of 2Gbps. In comparison, LTE deployed in 20MHz increments supports a theoretical maximum of 150Mbps and a practical capacity of just over 100Mbps. Given that a single stream of HDTV video requires about 7Mbps of throughput, a single LTE Advanced base station could support more than 275 simultaneous video streams (Exhibit 17). With small cell prices under $20,000, these low power micro base stations could be deployed to light poles and building rooftops to deliver capacity into individual neighborhoods. Since deploying a base station brings all households within range of its signal to its addressable market and can be redeployed as needed, wireless broadband would not be subject to the upfront risk of installing cabling to a neighborhood ahead of signing on subscribers. With current residential broadband prices rising above the $50/month level and cable MSOs saber rattling about future price hikes, there appears to be plenty of room for a wireless operator to attack.
Exh 17: Throughput Requirements by Application and Select Compression Technologies
LTE Advanced small cell overlay networks in residential neighborhoods are more than realistic over the next several years. Sprint, with Clearwire’s robust spectrum holdings, would seem a ripe candidate to pursue such a strategy, particularly if it can secure development support from a cash-rich internet partner. Dish Network’s free-agent spectrum is another option for a residentially focused service, as would a slice of the 700MHz band due to be auctioned.
Winners and Losers
We believe that the shifting competitive tide of the US wireless market will benefit both of the Sprint/Softbank/Clearwire and T-Mobile/MetroPCs combinations. The deterioration of the wireless oligopoly will also benefit network suppliers. We like the tower companies (American Tower, Crown Castle, and SBA), and network test companies (Agilent, Aeroflex, JDS Uniphase, and Ixia), but remain cautious on base station suppliers (Ericsson, Alcatel Lucent, Nokia) given the relentless global price pressure in that market. We also believe wireless industry trends will work to the great benefit of cloud-based application companies, such as Google, Amazon, Apple, Facebook and Microsoft. Of course, the biggest winners would be American consumers who would enjoy lower prices, fewer limits, higher speeds, and more choice (Exhibit 18).
Exh 18: The Winners and Losers
The obvious losers would be Verizon and AT&T, which could find their spectrum advantage ameliorated, forcing them to accept lower prices, and higher data limits, thus driving them to higher capital spending and lower profits. Similarly, we believe that the wireless threat to fixed residential broadband is real, putting cable operators (Comcast, Cablevision, TimeWarner Cable) under unexpected pressure.