When the FCC announced the November 18th release date for their long-awaited broadband mapping update, reflecting location-specific broadband availability as of June 2022, analysts at MoffettNathanson thought it would contain information on how many of U.S. homes have access to broadband and how many are too rural and are therefore unserved. However, that FCC release didn’t offer the numbers they needed, and the market research fim didn’t
have the necessary information to calculate it themselves.
In the underlying FCC datasets, which are provided for public download, each location served by a given technology or provider is a separate entry. One location is equivalent to one street address. But many street addresses in the U.S. correspond to multiple living units, and the number of units per location is not publicly available (the location fabric used by the FCC was contracted to a third party, CostQuest Associates, and that fabric is provided only to the FCC, broadband providers, state/local government entities, and select other interested parties). With approximately 31% of residences in multifamily homes, according to a 2019 survey by the Census Bureau, the number of units per location was, as of the November 18th release, a crucial missing piece for any meaningful coverage analysis we could do on our own.
Principal Analyst Craig Moffett wrote:
The FCC’s new maps of broadband availability can tell us coverage for residential locations or business locations, but not the combined total. The companies we cover sometimes break out residential and commercial, but not always. [As an aside, about half of small businesses in the U.S. are actually operated out of peoples’ homes, but hopefully this, at least, doesn’t introduce further distortion, since we are presumably still seeing just one subscription for one location]. So we’ll do our best to make sure we’re matching numerator and denominator by specifying whether we’re looking at all locations or residential locations only.
The FCC’s coverage data also doesn’t distinguish between occupied and vacant units. For our calculation of penetration, we’d want to exclude most vacant units, since vacant units don’t need broadband. Excluding all vacant units likely understates the denominator, though; for example, some second homes (which are treated as vacant) may have year-round broadband subscriptions. The best we can do is assume the coverage of total units is the same as the coverage of occupied units, and that vacant units with broadband subscriptions are negligible.
The FCC does report service coverage for satellite and fixed wireless. But some of those FWA subscribers are in areas where there’s no access to wired broadband, while others are in areas where wired broadband is available. Naturally, the companies won’t tell us how many of each there are. So we’ll just have to leave them all out. We’ll focus just on the availability of wired broadband.
Editor’s Note: The FCC broadband map for my address show a Licensed Fixed Wireless operator serves my condo. It’s California Internet with symmetrical 1G upstream/1G downstream. Also, there are two Satellite providers – Hughes Network Systems, LLC 25M/3M and Space X 350M/40M. Wired internet is available from AT&T and Comcast.
We’d really want to know how many DSL subscribers are in each of those different cohorts. But the
companies we cover don’t report how many of their DSL subscribers are in areas where there is
also a cable or fiber operator, and how many are in areas where DSL is the only option. The first
group is at risk. The second group is not. So, we’ll just have to include all DSL.
According to the FCC’s current estimates, wired broadband (defined as anything over 200 kbps downstream and 200 kbps upstream) was available to 93.7% of residential units in America as of June 30, 2022. Again, we don’t know the percentage of occupied housing units with wired broadband available, but let’s assume it’s the same. And we don’t know the number of residential units in the location fabric, so we’ll use the Census Bureau’s estimate of 128.1M occupied housing units in the U.S. Given these assumptions, we estimate wired broadband was available to around 120.0M occupied housing units as of June 30, 2022. With, by our count, an estimated 104.9M residential wired broadband subscriptions in America in Q2 2022 – again, including DSL, and sometimes including commercial as well as residential subscribers – that translates into penetration of 87.4% of broadband-available homes. We estimate that 81.5% of all households subscribe to wired broadband.
The goal for the FCC is to create maps that are not frozen in time but instead become living and breathing reflections of a dynamic marketplace. The new maps are subject to a public challenge process, enabling interested parties, including operators, local governments, and even individual would-be subscribers, to dispute reported availability. Challenges will eventually be part of a routine updating process. Indeed, the maps released in November were the product of what had already been a months-long initial challenge process. The maps are, again, a critical input to distribution of $42.5 billion of funds earmarked for rural broadband by the JOBS/Infrastructure Act. The National Telecommunications and Information Administration (NTIA) is required by law to use the FCC’s new map to distribute those funds in what is referred to as the Broadband Equity, Access, and Deployment (BEAD) Program, something they have committed to do by June 2023. They are likely to begin that process almost immediately, based on the number of unserved locations in each state, although NTIA chief Alan Davidson has said they will wait for the FCC to release the second version of its coverage map, later this year, before they actually begin to disburse those funds.
The network operators themselves, including the cable operators in particular, will in our view be major participants in the BEAD process, bidding aggressively to bring broadband to unserved census blocks on the periphery of their current franchise areas.
Meanwhile, Leichtman Research Group indicates that 90 per cent of U.S. households get an Internet service at home, compared to 84 per cent in 2017, and 74 per cent in 2007. Broadband accounts for 99 per cent of households with an Internet service at home, and 89 per cent of all households get a broadband Internet service – an increase from 82 per cent in 2017, and 53 per cent in 2007.
These findings are based on a survey of 1,910 households from throughout the United States and are part of a new LRG study, Broadband Internet in the U.S. 2022. This is LRG’s twentieth annual study on this topic.
Other related findings include:
- Individuals ages 65+ account for 34% of those that do not get an Internet service at home
- 56% of broadband subscribers are very satisfied (8-10 on a 1-10 scale) with their Internet service at home, while 6% are not satisfied (1-3).
- 44% of broadband subscribers do not know the download speed of their service – compared to 60% in 2017
- 61% reporting Internet speeds of >100 Mbps are very satisfied with their service, compared to 41% with speeds <50 Mbps, and 57% that do not know their speed
- 40% of broadband households get a bundle of services from a single provider – compared to 64% in 2017, and 78% in 2012
- 59% of adults with an Internet service at home watch video online daily – compared to 59% in 2020, 43% in 2017, and 17% in 2012
“The percentage of households getting an Internet service at home, including high-speed broadband, is higher than in any previous year,” said Bruce Leichtman, president and principal analyst for Leichtman Research Group, Inc. “Computer usage and knowledge remain the foundation for Internet services in the home. Among those that do not get an Internet service at home, 58% also do not use a computer at home..”
From two recent research reports to clients, MoffettNathanson chief analyst Craig Moffett wrote:
There is no question that there will be a great deal of new fiber deployed in the U.S. But we expect it will be considerably less than current worst-case scenarios for two reasons.
- There simply isn’t sufficient labor availability for all operators to meet the projections they’ve set forth (this issue will be significantly exacerbated by the upcoming rural Broadband Equity, Access, and Deployment (BEAD) program, which will introduce a dramatic new source of labor demand).
- The expected return from fiber overbuilds will be disappointing, in our view, both because deployment costs (including the cost of capital) have risen sharply, and because expected densities of available markets are falling sharply.
We are skeptical about the returns that will be generated by fiber builds, as costs are rising and densities are falling. The spiraling costs of fiber deployment also make it likely that there will be upward, not downward, pressure on broadband ARPU in competitive markets, as overbuilders scramble to cost-justify not only their existing projects, but, perhaps more importantly, the projects on which they have not yet broken ground (and which, without a more generous ARPU assumption, can no longer be return-justified). Craig had argued earlier this year that the fiber buildout bubble may pop.
Wireless operators have an enormous cost advantage in offering fixed wireless access (FWA) service on preexisting network facilities; the marginal cost of offering FWA is zero if it is simply using excess capacity. The capacity available for such a strategy is relatively limited, making the strategic leverage of FWA relatively limited as well. Cable operators have a smaller, but still significant, cost advantage in offering wireless services that can offload at least some of their traffic onto existing infrastructure. And unlike wireless operators offering FWA, their capacity to do so is unlimited.
Almost no telecom investor with whom we have spoken views FWA as an important part of the story for the companies that actually offer it. Investors seem to have already come to the view (for the wireless operators, at least) that FWA is at best a costly sideline in rural markets. Longer term, the bigger threat to cable broadband is likely fiber rather than fixed wireless, Moffett said. But even with that, the analyst seems to be less concerned that cable operators will overspend on fiber or that overbuilders will present more competition.
The convergence arguments for fiber to the home (FTTH) are arguably even weaker. As we’ve pointed out often, AT&T’s wireline footprint covers but 45% or so of the U.S. (by population), and of that, just a third is wired for fiber. In total, then, AT&T can deliver a bundled solution to just 15% or so of the population. In our view, a strategy (bundling) that “works” in 15% of the country isn’t a strategy.
We certainly aren’t convinced that the U.S. market will be fundamentally shaped by convergence. But if it is, the cable operators, not the telcos, are positioned to benefit.
by Craig Moffett of MoffettNathanson (edited by Alan J Weissberger)
Cable company’s 98% share of the wireless industry subscriber growth in Q1-2022 is a little known fact. And that does not include the free WiFi they offer to their customers, e.g. Xfinity WiFi and CableWiFi® (created through a collaboration of U.S. Cable and Internet Service Providers including Cox Communications, Optimum, Spectrum, and XFINITY. It allows each other’s eligible Internet customers free access to a collective network of more than 500,000 WiFi hotspots across the nation).
Until Q1-2022, Cable’s gains were almost exclusively from Comcast and Charter. Altice has now renegotiated its contract with T-Mobile, and they have moved to pricing that is even more aggressive than Comcast and Charter. [Interestingly, Altice’s contract allows Altice to name T-Mobile in their advertising as the underlying network, a contract term we’ve not seen before.]
Cox Communications, the nation’s third largest cable operator, is poised to join Cable’s ranks in offering wireless service, as well. The company won a Delaware Supreme Court decision in March, reversing a lower court decision that had previously upended their launch plans by finding they were bound to launch using T-Mobile’s instead of Verizon’s network, even if doing so was under less favorable contract terms. [They have not yet announced precise timing for their expected wireless relaunch.]
The pending addition of a wireless offering from Cox, and the more aggressive posture from Altice, will certainly compound the pressure Cable is putting on cellular telcos (e.g. Verizon, AT&T, T-Mobile, US Cellular, etc).
Cable’s 703K combined net additions were their best ever, and they have grown their subscriber base to 8.4M customers incredibly quickly. But those 8.4M subscribers still represent less than 3% market share of the U.S. market. They have a very long runway ahead.
Cable has achieved these gains without offering handset subsidies, something that seems inevitable sometime before the end of the year (Cable’s originally-BYOD subscribers will eventually demand new devices). Handset subsides from Cable, if and when they come, will only put more pressure on the cellular telco incumbents.
It is through this lens that one must view Verizon’s attempt to lead the industry to higher pricing [1.]. Subscriber growth is slowing. Cable’s share gains are accelerating. Cable has made clear that they do not plan to increase pricing. Nor does the industry price leader, T-Mobile. And Cable’s promotionality is likely to increase. That is a very tough backdrop against which to assume that price increases will “stick.”
Note 1. Verizon’s price increase, which will drop directly to the bottom line, will increase Verizon’s
service revenue and EBITDA by roughly $750M over the balance of the year, and by approximately $1.5B for next year, neither of which was contemplated in their previous guidance.
After accounting for 3G terminations, AT&T’s growth fell to just 360K net additions, leaving T-Mobile once again the industry’s fastest growing cellular telco. Not only is T-Mobile taking the industry’s largest share of gross additions – the best measure of customer choice – their churn rate is falling faster than any in the industry, as well, as they complete the transition of Sprint subscribers to their T-Mobile Magenta network. T-Mobile’s falling churn rate starves the industry gross add pool of what had been a critical source of “supply.”
The company is now most of the way through their migration of Sprint customers, and they have repeatedly suggested that churn on their Magenta network is the lowest in the industry, suggesting that churn should continue to fall, even if at a slightly slower pace going forward.
Only T-Mobile is growing ARPU at the moment, as more customers opt for higher value plans (Magenta Max). In contrast to the positive ARPU trend at T-Mobile, ARPU growth has been negative for eight straight quarters at AT&T (with the moderation in the rate of decline in Q1 largely attributable to the aforementioned extension of customer lives, which reduced amortization of historical promotional subsidies, and an easier comp against the same period last year).
ARPU growth at Verizon is not only negative, it is accelerating downward. For Verizon to post both negative subscriber growth and ARPU growth is a shock, and it points to the challenge facing the industry in getting ARPU increases to stick. Absent their wholesale contract with Cable, Verizon’s anemic 1.5% service revenue growth would be close to zero.
Verizon’s price increase comes at a time when industry unit growth is slowing, and at a time when Cable’s market share gains are accelerating both at the gross addition and net addition level. Without broad industry buy-in, and with subscribers looking harder to come by, we find it unlikely that Verizon’s price increases – even if AT&T does initially follow – will “stick.”
If wireless industry growth continues to decelerate, and Cable’s growth rate remains high, Cable’s share of growth will remain elevated, and the wireless industry will increasingly resemble a zero sum game for the Big Three incumbents, where one player’s gain (T-Mobile’s) will necessarily be another’s (Verizon’s and AT&T’s) loss. Huge losses at Dish Network’s Boost unit, and losses at U.S. Cellular, have helped soften the blow, but they are only so big. The pressure of falling industry growth and falling market share unavoidably falls on the cellular telco incumbents.
Wireless Q1 2022: The Elephant in the Room, MoffettNathanson report to clients
According to a new comprehensive, market research report from MoffettNathanson (written by our colleague Craig Moffett), Q4 2021 broadband growth, at +3.3%, “remains relatively robust,” and above pre-pandemic levels of about +2.8%.
Meanwhile, the U.S. fixed wireless access (FWA market) captured ~ 38% share of broadband industry net adds in the fourth quarter of 2021. Approximately half of Verizon’s FWA customers are coming from commercial accounts, T-Mobile has indicated that about half its FWA customers are coming from former cable Internet subscribers. FWA’s strong Q4 showing left cable’s flow share at just 66%, about the same as cable’s share of installed US broadband households. “In other words, Cable likely neither gained nor lost share during the quarter, and instead merely treaded water,” Moffett noted. FWA “has gone from low-level background noise to suddenly a major force, with Verizon and T-Mobile alone capturing more than 300K FWA subscribers in the fourth quarter,” Craig noted. However, he isn’t sure that wireless network operators will allocate enough total bandwidth capacity for FWA to fully scale.
In 2020, a year that witnessed a surge in broadband subs as millions worked and schooled from home, the growth rate spiked to 5%. Here’s a snapshot of the broadband subscriber metrics per sector for Q4 2021:
|Q4 2021 Gain/Loss
|Q4 2020 Gain/Loss
|Year-on-Year Growth %
|* Verizon and T-Mobile only
U.S. broadband ended 2021 with a penetration of 84% among all occupied households. According to US Census Bureau data, new household formation, a vital growth driver for broadband, added just 104,000 to the occupied housing stock in Q4 2021, versus +427,000 in the year-ago period. Moffett said the “inescapable conclusion” is that growth rates will continue to slow, and that over time virtually all growth will have to stem from new household formation.
Factoring in competition and other elements impacting the broadband market, MoffettNathanson also adjusted its subscriber forecasts for several cable operators and telcos out to 2026. Here’s how those adjustments, which do not include any potential incremental growth from participation in government subsidy programs, look like for 2022:
- Comcast: Adding 948,000 subs, versus prior forecast of +1.25 million
- Charter: Adding 958,000 subs, versus prior forecast of +1.22 million
- Cable One: Adding 39,000, versus prior forecast of +48,000
- Verizon: Adding 241,000, versus prior forecast of +302,000
- AT&T: Adding 136,000, versus prior forecast of +60,000
Are we witnessing a fiber bubble?
“The market’s embrace of long-dated fiber projects rests on four critical assumptions. First, that the cost-per-home to deploy fiber will remain low. Second, that fiber’s eventual penetration rates will be high. Third, that these penetration gains can be achieved even at relatively high ARPUs. And fourth, that the capital to fund these projects remains cheap and plentiful.
None of these assumptions are clear cut. For example, there is an obvious risk that all the jostling for fiber deployment labor and equipment will push labor and construction costs higher. More pointedly, we think there is a sorely underappreciated risk that the pool of attractive deployment geographies – sufficiently dense communities, preferably with aerial infrastructure – will be exhausted long before promised buildouts have been completed.
Revenue assumptions, too, demand scrutiny. Cable operators are increasingly relying on bundled discounts of broadband-plus-wireless to protect their market share. What if the strategy works, even a little bit? And curiously, the market’s infatuation with fiber overbuilds comes at a time when cable investors are growing increasingly cautious about the impact of fixed wireless. Won’t fixed wireless dent the prospects of new overbuilds just as much (or more) as those of the incumbents.”
Moffet estimates that about 30% of the U.S. population has been overbuilt by fiber over the past 20 years, and that the number is poised to rise as high as 60% over the next five years. But the big question is whether there’s enough labor and equipment to support this magnitude of expansion. “Our skepticism about the prospects for all of the fiber plans currently on the drawing board is not born of doubt that there is enough labor to build it all so much as it is that the cost of building will be driven higher by excess demand,” Moffett explained. “There are already widespread reports of labor shortages and attendant higher labor costs,” he added.
“The outlook for broadband growth for all the companies in our coverage, particularly the cable operators, is more uncertain than at any time in memory. IMarket share trends are also more uncertain that they have been in the past. Cable continues to take share from the telcos, but fixed wireless, as a new entrant, is now taking share from all players. Share shifts between the TelCos and cable operators are suppressed by low move rates, likely due in part to supply chain disruptions in the housing market. This is likely dampening cable growth rates. In at least some markets, returns will likely be well below the cost of capital,” Moffett forecasts.
U.S. Broadband: Are We Witnessing a Fiber Bubble? MoffetNathanson research note (clients and accredited journalists)