AT&T Finalizes FTTH Joint Venture, Joins a Growing List of ISPs Doing So

At some point in the future, we will be able to assess the extent to which multiple fiber-to-home and gigabit-per-second networks can survive in any single residential area, as we are reaching a point where some areas might have three to four suppliers of gigabit-per-second internet access.

The general issue is whether three or four providers is sustainable long term, as a general rule of thumb is that sustainable operation requires an installed customer base of at least 20 percent of passed locations, with operating costs scaled to match. 

And the number of suppliers is growing, including a number of joint ventures whose business models are based on reducing costs, boosting market share potential, or both. 

The Gigapower optical fiber joint venture between AT&T and BlackRock’s Diversified Infrastructure unit announced in December last year, is among the latest entrants. The venture has been finalized (the deal has closed), AT&T says. 

The joint venture is the latest to be formed by incumbent service providers in the U.S. market. Among the largest efforts is Nuvia, a joint venture between T-Mobile and Nokia, which aims to pass about 10 million locations by 2025. 

Atlanta, Charlotte, and Nashville are the areas where Nuvia expects to have its network operating by the end of 2023. It appears Nuvia is building first in “underserved” areas of those communities. 

Joint Ventures	Number of Locations
AT&T and BlackRock	1.5 million
Cox Communications and Google Fiber	2 million
CenturyLink and DigitalBridge	1.2 million
Frontier Communications and Google Fiber	2 million
T-Mobile and Nokia	10 million

One might argue about whether the strategy is to reach “underserved” neighborhoods or “most lucrative” neighborhoods where the demand is highest, irrespective of the number of existing ISPs serving those areas. 

The Buckhead area of Atlanta already is served by gigabit-per-second networks owned by Cox Communications and AT&T, with Google Fiber serving some parts of Buckhead as well. So Nuvia would be the fourth gigabit provider and the third FTTH provider. It is hard to say Buckhead actually is underserved. 

Nuvia Initial FTTH Build Areas
Nashville	Midtown, East Nashville, West Nashville, Sylvan Park, Green Hills, Belle Meade, Forest Hills, Brentwood, Franklin	100,000
Atlanta	Midtown, Buckhead, Downtown, Morningside, Virginia-Highland, Decatur, Sandy Springs, Alpharetta, Roswell	200,000
Charlotte	Uptown, Dilworth, Myers Park, South End, NoDa, Plaza Midwood, Ballantyne, Huntersville, Cornelius	150,000

Nuvia is T-Mobile’s first facilities-based effort to enter the fixed networks business as a FTTH supplier. 

Gigapower is intended to serve customers outside of AT&T’s traditional 21-state wireline service footprint. The original indications were that some 1.5 million new locations would be served by the wholesale network, open to other customers. 

In addition to Las Vegas, Gigapower now expects to expand beyond its previously announced fiber deployment in Mesa, to the Chandler and Gilbert areas of Arizona.also plans to build fiber in parts of Northeastern Pennsylvania (including Wilkes-Barre and Scranton) as well as parts of Alabama and Florida that are outside AT&T’s current service areas.

AT&T itself will be the initial anchor tenant. “Recently our first wholesale customer, AT&T, activated end-user customers on the Gigapower fiber network in Mesa,” AT&T Gigapower CEO Bill Hogg. 

Much of the current activity seems targeted at “tier two” cities.

Tier 1 Cities	Tier 2 Cities	Tier 1 Attributes	Tier 2 Attributes
New York City, Los Angeles, Chicago, Houston, Phoenix, Philadelphia, San Antonio, San Diego, Dallas, San Jose	Atlanta, Austin, Boston, Charlotte, Cincinnati, Cleveland, Denver, Detroit, Indianapolis, Jacksonville	Large population, major economic center, strong job market, high cost of living, diverse population, cultural and educational institutions, international airport	Smaller population, growing economy, moderate cost of living, less diverse population, fewer cultural and educational institutions, regional airport

Lumen Eliminates 60% of its Enterprise SKUs

Lumen Technologies says it eliminated more than 60 percent of its more than 12,000 enterprise SKUs (stock keeping units or products)  in less than a year. If one assumes there are about 100,000 to 120,000 total SKUs on tariff sheets globally, that implies a reduction of 60,000 to 72,000 SKUs should be possible, without damaging sales, revenues or profits.

In fact, according to Lumen, eliminating unused or legacy SKUs should reduce operating costs. 

But the existence of that many SKUs also illustrates the likely gap between revenue produced by a few leading products and revenue produced by almost everything else. Of course, no service provider ever reports revenue “by SKU.” Instead, SKUs get lumped together, typically showing a few big buckets such as home broadband or internet access, subscription video or voice services. 

Fixed and Mobile Service Provider Top Revenue Categories (Generic)
Fixed Network SKUs	% of Revenue	Mobile Service SKUs	% of Revenue
Internet Access	40%	Internet Access Data Plans	30%
Television Service	25%	Voice/Messaging Service	20%
Phone Service	15%	Mobile Hotspot Plans	15%
Other	10%	Other including phone sales	35%

The point is that a relatively small number of actual SKUs drive most of the revenue for any fixed network or mobile service provider, despite the existence of hefty catalogs. 

Mobile data roaming services, for example, might represent $8.6 billion in annual revenue for mobile service providers, on a global basis, in 2023. That likely means data roaming revenue, even if important, contributes far less than $1 billion in revenue for even the largest mobile operators.

Why Almost All Questions for Connectivity Providers are About Business Models

My interest in business models became very pronounced around 2000, as this chart by Benedict Evans might illustrate. If you were involved in “old media” (it was just “media” at the time), the growth of internet advertising, and the decline of magazine, newspaper, TV and radio advertising shows why business models became such an acute issue.

As somebody whose core business model for decades was “media,” this has been immediately relevant as a business challenge. 

The whole problem of “how do we make our money” and “how much can we make” has been a consistent issue for lots of people who run lots of companies, ever since the internet emerged. If you ran a media business at any point since about 2000, all other questions were subsidiary to the primary question of sustainable business models. 

Some of you with roots and history in the former telecom business know the challenge as well. Global statistics can be deceiving as more human beings start to use mobile phones and the internet. But in developed markets, there actually is less “communications” revenue being earned than there was in 2000, for example. 

That explains the absolute relevance of thinking and leadership around business models in the connectivity business. As with media, huge changes have taken place, and are taking place, in the sustainability of revenue and business models.

Consumer Satisfaction Plummets. Will Profits Follow?

According to conventional wisdom, customer satisfaction is correlated with profit margins. But not always. It appears that high inflation is dramatically hammering satisfaction scores, while profit margins are moving in the other direction, according to ACSI figures.

“The more recent downturn in customer satisfaction is worrisome,” say researchers at the American Customer Satisfaction Index. “It is often the case that a change in customer satisfaction precedes a change in profitability.”

“Since ACSI is down sharply, chances are that corporate profitability will follow suit,” ACSI says. And that is a sign of “malfunctioning markets.” 

The negative correlation between profits and satisfaction “is likely to happen in malfunctioning markets with supply problems for products as well as labor, demand greater than supply and where inflation is, at least in part, caused by collusion and price gouging,” the ACSI says. 

Internet service providers traditionally rank at the bottom of industry scores for customer satisfaction , along with linear video subscriptions. Mobile scores are higher and  fixed network voice scores, traditionally low, have stabilized, probably because disaffected customers have abandoned the service, so the remaining customers arguably are those who still see value in the service, as illustrated by eMarketer data. 

ISPs, as a group, rank about 20 points lower, on average, as shown by ACSI figures. 

As a rule, physical products industries tend to rank higher than products sold as subscriptions. 

Perhaps consumers are reminded every month when they have to pay anew for their subscription products of the value-price relationships. On the other hand, that “pay as you go” reality does not similarly seem to damage supermarket scores, which tend to be high. 

Whatever the reason, I cannot remember a time in many decades when communications services did not rank at the bottom of industry satisfaction scores.

FTTH Payback Models are Way More Complex These Days

If you are the sort of person who has followed the investment case for fiber to the home over a long period, you are well aware that payback periods are moderately long. You might also profess a reasonable belief that the payback analysis has gotten more complicated of late. 

Among the reasons: more of the payback might come from business customers, arguably less from consumer customers. More of the payback might come from attributed value for owners of mobile access services and less from services for customers of the fixed network. 

There are new investors in digital infrastructure whose estimations are financial rather than as operators of such assets. Such financial investors might not be as riveted on longer-term operating value but rather a chance to boost the valuation of assets before selling to longer-term investors. 

All of that requires a more-complicated analysis than was the case 30 years ago, when the analysis might generally be more simple: expected revenue gains from consumer services.  

It no longer is so clear that a quick, high level analysis still hinges largely on consumer revenue per account, for example, even if that might represent the principal direct revenue sensitivity.  

For example, where it once was possible to estimate consumer account revenue at perhaps $120 a month, as was the case for internet service providers selling home broadband plus entertainment video plus voice services, many ISPs now report that typical recurring home broadband revenue is more on the order of $50 to $70 per month. 

So revenue sensitivity also is potentially different for firms with other contributors, whether that is the value for mobile revenue generation, business services or wholesale upside. Pure-play home broadband ISPs almost always must contend with lower revenue per account than multi-play service providers with meaningful voice, video or other revenue drivers, including incremental revenue from mobile services that the FTTH investment supports. 

Some multi-play providers, including cable companies, also must contend with a major platform change similar in magnitude to a copper access telco upgrading to FTTH, though always with higher per-account revenue assumptions than a pure-play home broadband supplier. 

Firm	Breakeven Years	Investment per Home Broadband Account, Triple Play or Dual Play	Revenue per Account
Shentel	5-7	$1,500-$2,000	$100-$150
Google Fiber	3-5	$1,000-$1,500	$70-$100
AT&T	7-10	$1,500-$2,000	$100-$150
Comcast	10-15	$2,000-$2,500	$120-$170
Private Equity	7-10	$2,000-$2,500	$120-$170

If we assume that home broadband is the sole revenue driver, and that customers will routinely buy service plans at the upper range of available service plans (gigabit per second or multi-gigabit service plants), the payback models are a little more stringent.

Firm	Breakeven Years	Investment per Home Broadband Account	Revenue per Account
Shentel	10-12	$1,500-$2,000	$70-$100
Google Fiber	7-9	$1,000-$1,500	$50-$70
AT&T	12-15	$1,500-$2,000	$70-$100
Comcast	15-18	$2,000-$2,500	$90-$120
Private Equity Investor	10-12	$2,000-$2,500	$90-$120

The payback models are worse if the typical ISP has only home broadband as the primary revenue source and recurring revenues are at the lower end of what investors expect, about $50 per customer account, to start with, growing over time to higher levels over time, allowing for some annual increases in revenue per account. 

Firm	Breakeven Years	Investment per Home Broadband Account	Revenue per Account
Shentel	15-17	$1,500-$2,000	$50-$70
Google Fiber	10-12	$1,000-$1,500	$30-$50
AT&T	17-19	$1,500-$2,000	$50-$70
Comcast	20-22	$2,000-$2,500	$70-$90
Private Equity	15-17	$2,000-$2,500	$70-$90

The payback case might be longer if one has to include borrowing costs to acquire assets and then invest in FTTH. And one might argue that no private equity firm will hold assets for the full period required to earn an expected return. 

Among the obvious other imponderables are the differences in asset valuation that might be the driver, not the payback period or the typical recurring revenue, but mainly the arbitrage on assets whose value is boosted. 

Competition, demand assumptions and therefore revenue-per-account expectations are different. The implied value for supporting core mobile operations now is a factor. Business revenue arguably is more dynamic. 

And investment objectives are a new issue for some classes of investors. 

The upshot is that FTTH payback models are much more complex than was the case 30 years ago.