C-Band Prices are Not the Full Story

What do prices for U.S. C-band spectrum tell you about the value of mid-band spectrum? At a strategic level, the C-band arguably will be as foundational for 5G as the AWS spectrum was for 4G. 

But much else has to be interpreted. Most reports focus simply on the amount of revenue raised in the auction (more than $80 billion) and the implications for mobile service provider balance sheets.

The total amounts spent are important, but also misleading, as the C-band auction also involved the largest single offering of low-band or mid-band spectrum, ever. The 280 MHz of capacity was easily four times the amount of AWS spectrum that proved to be foundational for 4G networks, and four times the amount of the Citizens Broadband Radio Service spectrum at 3.5 GHz sold in 2020. 

To be sure, the millimeter auctions held so far, and any to follow, represent more raw capacity, balanced by signal propagation issues that limit coverage. But millimeter wave spectrum has to be evaluated on different metrics than the low-band and mid-band spectrum.

Capacity might be an order of magnitude or more higher than mid-band spectrum and at least a couple of orders of magnitude more than low-band. That does not mean bidders will pay an order of magnitude more for that spectrum.

Looking at the AWS-3 auction, the recent Citizens Broadband Radio Service and the C-band auction, AWS-3 made 65 MHz of capacity available. The CBRS auction held in 2020 represented about 70 MHz of licensed capacity. The C-band auction was for 280 MHz of capacity.

Price hinges on value, and the CBRS auction represents assets with the most restrictions (where it can be used, power limits). The nationwide CBRS average price per MHz-POP across all categories in the auction was $0.217 price per MHz-POP.

The AWS-3 auction in 2015 had a national average price per MHz-POP of $2.20. 

The C-band auction had an average national price of $0.94 per MHz-POP. After expected clearing costs, the average price might reach $1.09 per MHz-POP. 

source: Sasha Javid

Also, some of the C-band spectrum blocks were considered more valuable because the spectrum could be cleared and put to commercial use sooner. The Block A frequencies--100 MHz in five 20-MHz sub-blocks in the 3.7- to 3.8 GHz range--are expected to be cleared for use by December 2021.

The B Block--100 MHz of five 20-MHz sub-blocks, will take perhaps a year longer to clear. The C block--80 MHz of 20-MHz channels--might also take a year longer to clear. 

So A Block spectrum was expected to sell at higher prices, and did. 

According to analysis by Sasha Javid, COO of BitPath, the average price per MHz-POP for a Category A license was about $1.31, while that for BC licenses was $1.18 and $0.55 respectively.

Compare that to past spectrum transactions of mobile capacity (auction or asset purchase in secondary markets), where millimeter wave assets have sold for as little as $0.014 per MHz-POP, while some AWS-3 spectrum has sold for as much as $2.73 per MHz-POP. 

source: Summit Ridge Group

The C-band auction will be expensive for the winners, in part because so much capacity was purchased: volume matters for spectrum purchases as for most other commodities. But the price per MHz-POP will be substantially lower than for the crucial AWS-3 spectrum, which arguably was foundational for 4G networks. 

For AT&T, Verizon and arguably a few others, C-band assets are essential to maintain relevance in 5G services. It is a cost of being in business, in other words. 

So the purchases are essentially a requirement for sustaining the business--table stakes--and less a means of gaining business advantage against competitors. Many fiber-to-the-home investments have similar character.

How Much Business Revenue Will 5G Drive?

Most observers now believe 5G revenue growth will be disproportionately driven by new use cases and value for enterprise applications. In large part, that is because businesses are expected to be the buyers of edge computing, network slicing and internet of things sensor applications. 

That noted, consumer revenues in Europe, for example, are stable at about 68 percent of total revenue (mobile and fixed). In some markets, business revenues could range from 30 percent to 40 percent of total. Smaller providers might generate only about 15 percent of total revenue from business segments. 

The implications are clear enough. It will be harder to move the revenue needle if revenue growth is driven by enterprise and business applications and use cases, compared to consumer use cases, simply because of the installed base of accounts. 

source: ETNO, Analysys Mason

Adding to the near-term possibilities is the fall in business revenues caused by Covid-19 economic shutdowns, which will crimp enterprise revenue and destroy as much as 60 percent of the smaller business customer base, permanently, at least for those small businesses forced to cease operations. 

To be sure, business average revenue per account is higher than consumer ARPU, on either mobile or fixed networks. In virtually all markets, the differential in the mobile business has narrowed substantially over the last decade.

But mobile and fixed businesses still derive the majority of revenue from consumer services. At least for the near term, it is likely that business bankruptcies will inhibit the return to 2019 levels of business customer revenues. 

source: ETNO, Analysys Mason

Optimists about 5G business customer revenues expect revenue lift from new use cases related to internet of things or edge computing, for example. Evidence from the value-added services beyond connectivity might shed some light on reasonable expectations. 

European telecom provider share of the business services market in Europe has fallen from 16.2 percent in 2014 to 14 percent in 2019, ETNO says. Those services, such as desktop management, hosting and colocation, security, public cloud services or unified communications, mostly are provided by third parties. 

source: ETNO, Analysys Mason

If mobile operators and other connectivity providers are as successful with value-added apps and services as they have been with other non-connectivity services, they might expect to capture a few percent of the opportunities, best case. 

source: ETNO, Analysys Mason

ETNO expects connectivity provider IoT revenues to represent about 2.7 percent of total revenue by about 2027, for example. 

Connectivity alone generally represents less than 10 percent of total revenue from an IoT application, ETNO notes. In 2019, total IoT connectivity revenue in the ETNO countries reached EUR1.1 billion, about 0.4 percent of total telco revenues in the region and 1.2 percent of business segment connectivity revenues. 

Revenues might grow steadily to reach EUR2.9 billion by 2027. In 2019, IoT accounted for 0.9 percent of mobile service revenue in Europe, and by 2027 this is forecast to have risen to just 2.6 percent. 

Vehicle communications will likely provide the single biggest industry vertical revenue, ETNO says. 

source: ETNO, Analysys Mason

Will 5G Lead to Higher ARPU or ARPA? Yes and No.

There is substantial volatility where it comes to expectations about revenue lift from 5G. Some expect very little, if any, revenue lift from consumer services. “5G may not provide much uplift in revenue or profit,” said Dan Hays, Strategy& partner. “Our surveys show two thirds of consumers will pay nothing more for 5G.”

Others might argue that a small premium might be available for a time, as was the case for 4G when first introduced. Yet others might argue that “5G” revenue will be higher, but primarily because customers will get 5G as part of a more-costly service plan, such as an “unlimited usage” plan. 

Some also argue that with some moves to bundle value, 5G consumer service revenue could increase. Ericsson, as a key supplier of 5G infrastructure, naturally is in that camp. 

Though Ericsson argues that mobile operators will gain revenue even if they stick to business models focused on internet access and voice, they also believe an“active” approach will boost returns.

Ericsson, for example, points to entertainment video, live sports streaming, music, gaming (mobile and cloud), augmented and virtual reality (AR/VR) and consumer IoT services as areas where packaging with 5G could boost average revenue per account or average revenue per user.  

Ericsson refers to such opportunities as “core” opportunities because they are either established as consumer services already offered by service providers, or present clear opportunities for 5G service providers to operate as providers of their own services or sales channels for services from third-party providers. 

 source: Ericsson

Ericsson also says there are “adjacent digital services” including digital advertising, in-car entertainment and connectivity, in-venue digital services, and advanced digital services (such as the internet of senses) which will bring new senses online.

There perhaps are tertiary opportunities in mobile commerce, such as consumer revenues generated through payments for physical goods and services which will run over 5G networks. Gaining roles of this sort will be difficult, Ericsson notes. 

It is worth noting that the revenue upside from core digital products and services likely involves some degree of ownership of the content assets. Otherwise, the mobile service provider remains a seller of commodity internet access. 

Most are more sanguine about business use cases, though even there the magnitude of increase is volatile as well. 

It is possible that 5G average revenue per account could increase if mobile operators change the product: offering content bundles with 5G; or bundling 5G access with higher-priced plans, for example. 

One might argue that 5G is a feature of such plans, not the driver of higher ARPU or ARPA.

When Infrastructure Sharing Makes Most, Least Sense

Infrastructure sharing is more popular in some markets than others, but the logic is always the same: competitors save on capital investment when sharing the cost of building infrastructure. 

The trade off always revolves around control, ability to craft distinct value propositions and a shift of spending from capex to opex. 

In this typology by Strategy Analytics, the least amount of sharing typically is for all mobile contestants to lease tower capabilities, rather than owning towers. That offloads the real estate, but retains all the rest of the active radio infrastructure, spectrum assets and backhaul. 

source: Analysys Mason 

The greatest amount of sharing might happen when real estate, radio infrastructure, backhaul and possibly some spectrum rights are shared. 

Sharing almost always makes the most financial sense in rural areas, where revenue upside is limited and business models are typically most constrained. Sharing typically provides the least value when traffic volume, revenue and potential profits are highest, as in core urban areas. 

Under those circumstances, owner’s economics are important.

How 5G ARPU is Increased Indirectly

To the extent that there is revenue uplift from 5G, the issue for virtually all mobile service providers is how to accomplish that objective. The common way we ask the question is “whether consumers or businesses will pay more for 5G.” 

As always in the consumer segment of the business, the optimistic answer is “maybe.” The realistic answer, based on history, is that a small premium might be possible at first, as was the case for 4G in some markets, with the premium vanishing with fuller deployment and competitive responses by competitors, so long as the price comparison was “like for like.”

In other words, where the same usage plans were compared head to head, was there a price difference? Yes, in some cases, at least initially. At scale, the premium tends to vanish.

In the U.S. mobile market, T-Mobile has a firm policy of “no incremental price increases to use 5G.” All major service plans include use of the 5G network with no price increase. 

That tends to limit the usefulness of a 5G price increase. But 5G is expected to lift revenue for AT&T and Verizon, despite those limitations. How? 

So far, the revenue boost has been accomplished by enticing customers to buy higher priced plans not specifically because of 5G, but to get “unlimited usage.” The value proposition is not 5G per se but unlimited usage. 

In the context of the Covid-imposed changes in mobile data usage, that is ironic. Most workers and students have been staying at home, which means consumption has largely shifted to Wi-Fi, reducing overall mobile data consumption. 

So right at the point where mobile data usage is declining, service providers are pushing unlimited-usage plans. To be sure, the lockdowns and stay-at-home rules will be lifted. Patterns will tend to revert to the mean. 

But if the predictions of greater and permanent work from home come to pass, then the demand curve will be permanently shifted to some degree. There will be less demand for mobile data than before, in the same locations and times of day and days of week.

Some urban cells will have less demand than before, while some suburban cells might see an increase. But much more traffic will shift to Wi-Fi as well. So one has to wonder how successful the “shift customers to unlimited-usage plans” will be, longer term, as a way of incentivizing use of 5G. 

In fact, one might infer from Verizon’s fourth-quarter earnings report that the effect of unlimited usage adoption was to offset other negative trends in the consumer mobility market. “For the full year (2020), consumer wireless service revenue was $53.6 billion, relatively flat from 2019 levels,” Verizon noted. 

In other words, the way Verizon expects to achieve consumer account revenue lift is to switch customers to higher-price unlimited plans that, oh by the way, is how 5G is made available. 

“Our competitive position is based on compelling unlimited plans that provide choice to consumers, paired with the best devices and attractive promotions, all built on the country's best network,” said Hans Vestberg, Verizon CEO. 

“We have now more than 60 percent of our customers on unlimited,” Vestberg noted. “But more important, I mean if you look in the fourth quarter, 90 percent of the net adds took unlimited, and out of those, 55 percent took premium.”

“And the premium is, of course, where we are adding in the 5G,” he said.

3GPP Release 16 Integrated Access and Backhaul Will Lower Small Cell Deployment Costs

One business model issue many observers have raised is the need to create dense networks of small cells to support 5G millimeter wave spectrum, something that puts T-Mobile and AT&T at a disadvantage, with Verizon at an advantage, said FibeReality founder Mark Lutkowitz at a #PTC21 session on advanced technologies. 

Millimeter wave spectrum is the game changer for 5G, “if there is one,” said Benoit Felten, Diffraction Analysis founder. 

source: Qualcomm

Qualcomm, though, believes there are ways to create dense millimeter wave networks based on small cells without pulling new optical fiber to each small cell. 

But 3GPP Release 16-defined integrated access and backhaul allows 5G base stations to backhaul wirelessly to neighboring base stations using the same mmWave spectrum used to support retail 5G access.

IAB will enable lower-cost densification strategies in the near term, allowing mobile operators to quickly add new base stations to their networks before having to install new fiber to increase backhaul capacity later, Qualcomm argues.

Where 5G Millimeter Wave Makes Sense

With the caveat that total cost of ownership is difficult to calculate, and also tends to represent less than 30 percent of total cost, GSMA Intelligence has attempted to calculate the net present value of millimeter wave 5G networks. As always, take rates matter, as does density. The GSMA analysis uses 2025 as the analysis period end date. 

 source: GSMA Intelligence

Adding millimeter wave capabilities to a 3.5-GHz network has a positive financial outcome in dense urban areas in China when peak hour use of the millimeter wave network is at least five percent. In Europe the positive outcome happens when peak hour use of the millimeter wave network is at least 10 percent. 

Used to support fixed wireless, millimeter wave assets have highest value when there is strong data demand growth in China’s urban areas, Europe’s suburban areas and U.S. rural areas; and when there are specific localized areas of capacity undersupply. 

In indoor settings, millimeter wave networks are financially positive when the millimeter wave network carries between five percent and 20 percent of traffic. 

GSMA also assumes savings will be higher after 2025, as platform costs continue to decrease. 

Why 5G and Prior Next-Generation Networks were Necessary

Why 5G might be needed, when 4G handles most present use cases, is a fair question. Leaving aside for the moment new applications or use cases or even ever-increasing end user bandwidth demand, next-generation mobile networks always are needed because the networks themselves are built with specifications that eventually become limitations.

3G networks were designed to use channels 5 MHz wide, maximum. 4G was designed to use channels 20-MHz wide, maximum. 5G now uses 100-MHz channels, with a path to channels 2-GHz wide. 

Also, every generation of mobile networks are designed to run on specific frequencies. To add more spectrum necessarily means creating a new platform able to operate at the new frequencies. 

The network specifications therefore matter because all devices are designed with assumptions about network capabilities in mind. Though each new network adds capacity and better latency performance, performance eventually becomes an issue, as is the case for generations of personal computers or servers. 

In recent tests conducted by Signals Research Group, a 4K video delivered over a 4G network using channel bonding (two channels). The video buffered and froze about 79 percent of the time, while there were no impairments using 5G. 

Covid Recession Changes Very Little About Mobile Business

It is hard to see any long-term change in demand for mobility, smartphones or wireless in the wake of the Covid pandemic, anymore than it is possible to pin big shifts in demand to the internet bubble and recession of 2001 or the Great Recession of 2008. A shallow, short dip in revenues for service providers or phone manufacturers can be seen, but no fundamental reduction of demand after the recessions ended.

Economists might argue about the long-term effects of a major recession, though the near-term impact is normally a reduction in demand. A bigger issue is the effect on medium-term demand, and there is some evidence that a big recession depresses overall demand even eight to 10 years beyond the recession. 

But technology substitution and new product demand might have greater impact than any recession, no matter how severe, and that has arguably been true in the mobility business.

At the time of the internet investment bubble (2001), U.S. smartphone use was close to zero, as was use of “broadband” internet access services. Usage of both accelerated rapidly after the bubble burst. 

At the same time, fixed network voice demand began a sustained fall, suggesting that mobile voice had become the substitute for fixed network alternatives. Complicating matters was that a huge change in mobile voice pricing happened in 1998, when AT&T introduced Digital One Rate.

That new plan by AT&T Wireless drove rapid mobile adoption by consumers by revolutionizing the way consumers paid for--and how much they paid for--long distance calling.

The Digital One Rate pricing plan for mobile users priced all domestic calls at 10 cents a minute. That Digital One Rate plan effectively erased the distinction between local and long distance calling and provided a major incentive for consumers to buy mobile service as a way of controlling their long distance calling costs, which might have been 25 cents per minute or higher at the time.

In other cases a big recession does not necessarily seem to have changed behavior. Consider the smartphone. In 2001, when the internet bubble burst, smartphone adoption in the United States was very close to zero. Six years later, adoption still might have been in the 2.6 percent range. 

Not until about eight years after the internet bubble burst did smartphone adoption began its rapid ascent. That suggests the dot.com bubble did not cause the smartphone trend. 

source: Asymco 

Nor does the 2001 recession seem to have caused or accelerated the use of mobile phones in general. That trend was already strongly in place before the disruption. 

source: Harvard Business Review 

Likewise, many would note the more-rapid shift to cloud computing in the wake of the Great Recession of 2008, partly explainable by greater use of cloud-based conferencing apps, for example, and the need to support remote workforces, which, if nothing else, increased demand for cloud-based collaboration apps. 

Still, the turn of the century (with the internet bubble burst happening in 2001) also saw rapid adoption of many technologies even by “laggard” demographic groups. It may shock you, but an analysis by the Pew Research Center claims that senior adoption of “home broadband” was virtually zero in 2000. 

We can assume that huge adoption behavior was driven by the new perceived value of apps, content and services enabled by the internet. It is harder to view such growth as an outcome of the dotcom bubble. 

Social media exploded about 2008, the same time as the Great Recession. But it is impossible to conclude that the Great Recession “caused” the big upsurge in social media usage. 

source: Pew Research

C-Band Spectrum is 5G Table Stakes

Demand for C-band spectrum in the U.S. mobile market was high for a few reasons. Leaving aside for the moment millimeter wave assets, which are important for capacity gains, both Verizon and At&T were far behind T-Mobile in the amount of low-band spectrum--good for coverage--and mid-band spectrum that provides both reasonable coverage and noticeably higher data access speeds.

source: S&P Global

The C-band auction offered a huge trove of mid-band spectrum (280 MHz) that represents the  largest amount of spectrum ever awarded in a single low-band or mid-band auction. Millimeter wave assets are in a different category, as the capacity of frequencies above 6 GHz is an order of magnitude or two orders of magnitude greater than in the low-band, for reasons of physics.

To be sure, as when buying real estate, volume matters. That is why the total amount of C-band spectrum spending will set a record for licensed mobile auctions. 

But the C-band also is important for other strategic reasons. As the AWS-3 auction of about 70 MHz of spectrum was key for U.S. 4G, the C-band will underpin U.S. 5G. Aside from T-Mobile, all the other contestants essentially must have significant C-band assets or risk losing the 5G market share battle. Capacity is table stakes, and all the other service providers save T-Mobile needed C-band capacity. 

To be sure, the 70 MHz of new Citizens Broadband Radio Service shared spectrum will help. But C-band assets will be foundational in the 5G era, for the same reason capacity is important in the fixed network internet access business. If the market standard shifts to hundreds of megabits per second and then gigabit speeds, every major supplier has to be able to supply such speeds. 

It is simply a matter of remaining market competitive.

Alphabet Shutting Down Loon

Alphabet’s Loon balloon-based internet access effort is shutting down. Facebook shut down Aquila, its unmanned aerial vehicle effort to provide internet access, in 2018. That leaves low earth orbit satellite constellations as the remaining “new platforms” being developed to supply internet access especially to remote and underserved areas. 

The fate of other high altitude pseudo satellite efforts is unclear. Zephyr might wind up as a military or commercial sensing platform or a disaster relief tool, more than a consumer internet access platform. 

source: TM Forum

The ultimate success of  LEO constellations also remains to be tested. The first wave crashed in the 1990s. 

All of the efforts--failed and yet to be determined--simply illustrate the huge cost and tough business models for access networks generally, and especially for networks capable of supporting users in remote areas, at lowish cost.

U.K. Mobile Customers Seem to Have Rational Expectations about 5G

U.K. mobile service consumers and businesses seem to have realistic expectations about 5G service, namely the extent of coverage. That noted, consumers and business users continue to believe data speed is more important than coverage, RootMetrics data suggests. 

source: RootMetrics

Interestingly, a RootMetrics satisfaction survey shows a metro area performance pattern that mimics that of a stable market. The predicted “stable competitive market” pattern would show market shares in a 4:2:1 structure. That is almost precisely what the RootMetrics survey of metro area mobile performance reflects. 

source: RootMetrics

Actual market shares do not match the pattern, however. 

In early 2021, EE leads with 38 percent market share, followed by O2 at 29 percent, then Vodafone at 21 percent, with Three at 12 percent share. Vodafone is the firm whose metro area scores diverge most from its predicted market share pattern, were the U.K. mobile service market competitively stable.

5G Monetization Might Often be Indirect

Monetization of 5G services might not be based on sheer speed, says Jefferson Wang, Accenture Global 5G Strategy Lead. In fact, in many cases, mobile operators will not be able to charge a premium for 5G, much as customers do not pay extra for 4G service. 

Instead, they arguably pay for larger data usage buckets, music or video streaming subscriptions, cloud data storage, with voice and text usage essentially a feature of the access service. 

In fact, other things 5G enables might be the actual value driver, not 5G itself. 5G might be seen as a way to enable immersive experiences (gaming, entertainment, training), security (factory automation, health), responsiveness (autonomous vehicles, public safety, machine control). 

Source: Jefferson Wang, Accenture

In many cases, 5G will not be the retail product, but a capability that enables a retail product. Consumers or businesses buying 5G fixed wireless for internet access will not be buying “5G.” Instead, they will buy “fast internet access.” 

Other customers might be paying indirectly for 5G, as when 5G enables edge computing, which in turn enables some crucial application or application feature. In any logistics or manufacturing setting where many sensors are required, 5G might be the indirect value that enables near-real-time analytics. 

The shift of value and revenue drivers away from raw performance is not unusual in the communications business. In the mobile business, coverage matters as much as data speed. 

At some point, PC manufacturers stopped competing on “ever faster clock speeds.” Fixed network broadband access providers arguably have not reached that point, as speeds and price still remain the marketing battle essentials. 

Mobile operators might, by the end of the 5G era, have stopped competing on “feeds and speeds” and begun competing on other values: use cases that require 5G, much as the value of a smartphone or a PC now is based on all the applications one can access using a device.

Was C-Band Spectrum a "Bargain?"

The potential issue with U.S. mobile operator spending on C-band spectrum has less to do with the aggregate spending than with the business model mobile operators confront. Spectrum costs of about $81 billion--up to perhaps $94 billion after clearing costs--represent a huge capital investment impact, to be sure. 

On the other hand, costs on a per-unit basis arguably are quite reasonable. The larger issue is capex investment in a business whose revenue growth rates are perhaps one percent a year to 1.5 percent a year. 

As a strategic matter, nobody contests the notion that C-band spectrum (and mid-band spectrum generally) will be foundational for 5G, as the AWS-3 spectrum was foundational for 4G. 

Despite the reliance on unlicensed spectrum, existing low-band, existing mid-band and millimeter wave capacity, virtually everyone agrees that contenders including AT&T, Verizon, the cable operators, Dish Network and, to a lesser extent, T-Mobile “needed” that spectrum to compete. 

Overall C-band prices were about $0.94 per MHz-POP. The foundational 4G AWS-3 auction featured prices of about $2.20 per MHz-POP, in comparison. The issue is volume. The C-band auction of 280 MHz of capacity was about four times the amount of AWS-3 spectrum. 

If C-band and AWS-3 represented beachfront property, then the C-band auction represented four times as much beachfront property to be acquired. 

Spectrum prices always are highly influenced by the most-recent sale of similar spectrum. So it is logical to compare C-band prices with Citizens Broadband Radio Service prices, auctioned in 2020. That auction had average overall prices of just $0.21.7 per MHz-POP. So it might be argued that the C-band assets were wildly more costly.

But there are key qualitative differences. CBRS represented a total of 70 MHz of capacity, but on a shared basis: the licenses are not exclusive, and feature transmit power and geographic restrictions. 

On one level, it can be argued that some firms--including AT&T and Verizon--simply did not have a choice: they had to acquire substantial C-band spectrum--whatever the cost--or risk their ability to compete in the 5G era. 

The larger issue is the impact the capex will have on revenue and alternative uses of that capital to promote growth, retire debt or buy back stock, in a market many would argue has saturated, with little opportunity for significant incremental revenue growth.

On the other hand, some argue new revenue sources contingent on adequate 5G bandwidth,  will develop, partially offsetting the risk of spectrum acquisition. Also, like it or not, mobile service providers must keep increasing the amount of bandwidth their customers can use, at roughly stable prices over time. 

That makes additional spectrum costs unenviable, but also quite necessary. There will have to be offsetting inputs--cost and revenue--to accommodate the C-band spending.