Stephenson

AT&T CEO Randall Stephenson on Tuesday told investors that AT&T will deploy a combination of fiber optics and 5G wireless and be able to sell a “true, high-speed internet network throughout the United States” within the next three to five years.

“In three to five years out, there will be a crossover point,” Stephenson told investors. “We go through this all the time in industry. 5G will cross over, performance wise, with what you’re seeing in home broadband. We’re seeing it in business now over our millimeter-wave spectrum. And there will be a place, it may be in five years, I think it could be as early as three, where 5G begins to actually have a crossover point in terms of performance with fiber. 5G can become the deployment mechanism for a lot of the broadband that we’re trying to hit today with fiber.”

Although the remarks sound like a broadband game changer, Stephenson has made this prediction before, most recently during an AT&T earnings call in January, 2019. Stephenson told investors he believed 5G will increasingly offer AT&T a choice of technology to deploy when offering broadband service to consumers and businesses. In high-cost scenarios, 5G could be that choice. In areas where fiber is already ubiquitous, fiber to the home service would be preferred.

Stephenson’s predictions about nationwide service will depend in part on the commercial success of millimeter wave 5G fixed home broadband, which will be required to satisfy broadband speed and capacity demands. Verizon Wireless has been offering fixed 5G in several markets with mixed results. The company’s early claims of robust coverage have been countered by Verizon’s own cautious customer qualification portal, which is more likely to deny availability of service to interested customers than offer it.

But Stephenson remains bullish about expanding broadband.

“So all things considered, over the next three to five years, [with a] continued push on fiber, 5G begins to scale in millimeter-wave, and my expectation is that we have a nationwide, true, high-speed internet network throughout the United States, [using] 5G or fiber,” Stephenson said.

Whether anything actually comes of this expansion project will depend entirely on how much money AT&T proposes to spend on it. Recently, AT&T has told investors to expect significant cuts in future investments as AT&T winds down its government-mandated fiber expansion to 14 million new locations as part of approval of the DirecTV merger-acquisition. In fact, AT&T’s biggest recent investments in home broadband are a result of those government mandates. AT&T has traditionally focused much of its spending on its wireless network, which is more profitable. For AT&T to deliver millimeter wave 5G, the company will need to spend billions on fiber optic expansion into neighborhoods where it will place many thousands of small cell antennas to deliver the service over the short distances millimeter waves propagate.

AT&T could sell a fixed 5G broadband service similar to Verizon Wireless, confine its network to mobile applications, or offer fixed wireless service to commercial and manufacturing users in selected areas. Or it could offer a combination of all the above. AT&T will also need to consider the implications of a fiber buildout outside of its current landline service area. Building fiber optic networks to provide backhaul connectivity to AT&T’s mobile network would not antagonize its competitors nearly as much as the introduction of residential fixed 5G wireless as a home broadband replacement. The competitive implications of that would be dramatic, especially in communities skipped by Verizon FiOS or stuck with DSL from under-investing independent telephone companies like CenturyLink, Frontier, and Windstream. Should AT&T start selling 300+ Mbps fixed 5G wireless in these territories, it would cause significant financial distress for the big three independent phone companies, and could trigger a competitive war with Verizon.

Wall Street is unlikely to be happy about AT&T proposing multi-billion dollar investments to launch a full-scale price war with other phone and cable companies. So do not be surprised if AT&T’s soaring rhetoric is replaced with limited, targeted deployments in urban areas, new housing developments, and business parks. It remains highly unlikely rural areas will benefit from AT&T’s definition of “nationwide,” because there is no Return on Investment formula that is likely to work deploying millimeter wave spectrum in rural areas without heavy government subsidies.

For now, AT&T may concentrate on its fiber buildout beyond the 14 million locations mandated by the DirecTV merger agreement. As Stephenson himself said, “When we put people on fiber, they do not churn.” AT&T has plenty of runway to grow its fiber to the home business because it attracts only about a 25 percent market share at present. Stephenson believes he can get that number closer to 50%. He can succeed by offering better service, at a lower price than what his cable competitors charge. Since 5G requires a massive fiber network to deploy small cells, there is nothing wrong with getting started early and then see where 5G shakes out in the months and years ahead.

This is the second part of a two-part series reflecting on Verizon’s 5G millimeter wave wireless home broadband service and how Wall Street complicates its potential. Be sure to read part one, “How a Wall Street Analyst Complicates AT&T and Verizon’s Upgrade and Investment Plans” for the full story.

“Put simply, the cost of building a second network is so high that its builder simply can’t earn a passable return based on the market share available to a second player,” Craig Moffett, an important telecom industry analyst working on behalf of Wall Street investors, argued over Verizon’s fiber to the home project dubbed FiOS. “Virtually every overbuilder, from telephone companies to competitive cable companies to municipalities, has learned this lesson the hard way; almost all such efforts have ended in bankruptcy. Verizon’s own FiOS network was an economic failure; there is no longer any debate about whether FiOS did or didn’t earn its cost of capital. It didn’t, and it wasn’t even close.”

Moffett’s philosophy about emerging broadband technology and competition is heavily influenced by his personal and professional belief that broadband competition is bad for business and investors. His distaste for Verizon FiOS, a plan to scrap old copper phone wiring in favor of fiber optics, was well-known across the industry and trade press. But Verizon kept going with the project under the leadership of then-CEO Ivan Seidenberg, who was a telephone man through and through. But by 2010, Seidenberg had decided to retire, and his successor, Lowell McAdam, had a very different perspective about Verizon’s future. McAdam spent almost his entire career from the early 1990s forward in the wireless business. In 2006, McAdam was named the chief operating officer and CEO of Verizon Wireless. When he succeeded Seidenberg in late 2010, Verizon had already announced it was winding down further FiOS expansion. That seemed to suit McAdam just fine, because under his leadership as CEO of Verizon, Verizon Wireless became the dominant focus of the company. Heavy investment in wireless continued, while Verizon’s landline network was allowed to deteriorate.

Moffett told his clients the end of FiOS expansion would be good news for cable companies because they would lose fewer subscribers as a result.

Verizon’s marketing machine carefully lays its business case for 5G home broadband

More than a decade later, Verizon’s decision to embark on another major technology upgrade requiring billions in new spending quickly raised eyebrows on Wall Street. This time, however, Verizon executives attempted to be better prepared to defend their 5G vision from the reflexive investor argument that it was too expensive and extravagant.

Moffett

“First, their fixed wireless broadband business will leverage investments that Verizon argues they will need to make anyway to support their wireless network,” Moffett wrote in a report to his clients, acknowledging Verizon’s claimed reasons for entering the wireless home broadband business. “Second, Verizon argues that it will be cheaper to connect homes wirelessly than it is to connect them with fiber, making it economic to deploy fixed wireless in markets where fiber to the home hasn’t been economically justifiable.”

Most of the expenses cited by Moffett relate to bringing fiber networks into neighborhoods to support the small cell technology Verizon is relying on for its 5G home broadband and mobile services.

Moffett also believes the only attractive market for 5G service will be in more upscale suburban rings around cities, not densely populated urban centers or rural areas. Moffett argues fiber providers are likely already providing service in urban areas and rural areas simply lack enough customers to justify the cost of either a fiber optic network or a small cell network. Ironically, that conclusion means the same suburban ring Moffett rejected 5-10 years ago as economically unsuitable for fiber service is now precisely the area Moffett argues is the only attractive market for fiber service, to bring 5G.

From a short-term results perspective, laying fiber optics is a costly proposition unlikely to return much revenue gain in a few short years. That reality has kept many investor-owned phone companies away from expensive network upgrades. These legacy telephone companies recognize they are going to continue to lose customers to faster technologies like cable, fiber, and perhaps, wireless. But managing an existing low-speed DSL business seems preferable to facing the wrath of investors upset over the prospect of shareholder dividends and share buybacks being curtailed to redirect money into a full-scale upgrade effort, even if it results in better returns and greater revenue a few years down the road.

Verizon is depending on its wireless division’s extremely high profitability to counter the usual objections to major upgrades, and by focusing on how 5G will enhance the wireless experience. It also benefits from media hype surrounding 5G technology, exciting some investors. But Verizon is also downplaying exactly what it will cost to lay fiber optic networks deep into neighborhoods to deliver it.

Moffett investigates Verizon’s first 5G city — Sacramento, Calif., and discovers alarming results

Moffett decided to bypass the traditional cost-benefit analysis of laying mile after mile of optical fiber and decided to test Verizon’s case for wireless 5G home broadband instead.

Six months after launch, Moffett investigated Verizon’s 5G millimeter wave network in Sacramento, examining how the service is initially performing. Moffett identified seven zip codes in Sacramento where service was most likely to be available, based on cell tower/small cell records. As of late February, Moffett found Sacramento had 391 Verizon small cells installed, with 273 used for millimeter wave 5G service (the rest are likely designed to bolster Verizon’s 4G LTE network).

Verizon has admitted small cell technology is vulnerable to distance, so Moffett relied on earlier purported claims of 5G coverage to limit the number of addresses to be sampled. Moffett’s team identified 45,000 out of 70,000 possible addresses, based on if those homes were located within a radius of 0.7 miles of a 5G small cell. Then, Moffett’s team devised a method of hitting Verizon’s 5G availability website with each of those 45,000 addresses to learn which ones Verizon qualified for 5G service.

The results, so far, are underwhelming:

• Only an average of 6% of the queried addresses were actually eligible to receive Verizon’s fixed wireless service. That could mean Verizon has installed 5G small cells, but some are not yet operational in all areas or the network is performing much worse than originally anticipated. Some zip codes did better than others, but not by much. The best offered just an 18% pre-qualified acceptance rate. Apparently Verizon’s qualification website also informs applicants if they already have service, which proved to be a good way of finding out how many addresses actually have signed up. Moffett claims only 3% of eligible customers have decided to subscribe to Verizon’s 5G home broadband service so far.
• Coverage appears to a problem. As Moffett checked addresses further away from each small cell, more and more were deemed ineligible for service. In fact, despite Verizon’s claims that its 5G signal reached customers more than 1,900 feet away, the company’s own website refused to actually sell service to customers that far away. Moffett found subscribers were deemed ineligible for service as little as 400 feet away from a small cell. At that distance, less than 50% of checked addresses could sign up. For those 700 feet or more away, almost no addresses were qualified for service.

With those results, Moffett was able to extrapolate some important numbers about how much Verizon’s infrastructure is being utilized:

• Each small cell serves approximately 27 eligible addresses.
• Verizon’s 5G home broadband has a 0.1% market share in Sacramento.
• Excluding areas where multi-dwelling properties dominate, Verizon has achieved a penetration of roughly one subscribed single-family home per 1.5 5G small cell.

“Our findings in Sacramento — limited coverage, low penetration — preliminary though they may be, suggest that earning an attractive return will be challenging, at best,” Moffett concluded.

Because Verizon has attracted so few subscribers thus far, the total cost per connected home for 5G wireless service could far exceed what it would cost to just lay down fiber to the home service to each customer, which might actually give Verizon more business.

“Our analysis suggests that costs will likely be much higher (that is, cell radii appear smaller) and penetration rates lower than initially expected,” the report explained. “If those patterns are indicative of what is to come in a broader rollout, it would mean a much higher cost per connected home, and therefore much lower returns on capital, than what might have been expected from Verizon’s advance billing.”

If Moffett’s estimate of 27 residences served per small cell was proven true, Verizon would have to deploy well over five million small cells to deliver 5G wireless service across America.

Verizon’s choice of cities to launch its 5G millimeter wave network may be partly designed to test the differences in topology, building density and foliage levels, and there may be dramatic differences between Houston, Sacramento, Indianapolis, and Los Angeles.

Moffett’s overall conclusion is that should Verizon move forward on rolling out 5G wireless home broadband to around 25% of the country, as it planned, reaching those 30 million homes “will take a very, very long time, and it will cost a great deal of money.”

Verizon, the country’s leading provider of millimeter wave 5G wireless broadband, is promising to expand service nationwide, but admits it will only service urban areas where the economics of small cell/fiber network infrastructure makes economic sense.

At the Mobile World Congress conference in Barcelona, Spain, Verizon’s vice president of technology planning told PC that when it launches its mobile 5G network later this spring, home wireless internet service will come along for the ride.

“It is one network, based on 5G, supporting multiple use cases,” Verizon’s Adam Koeppe said. “Enterprise, small/medium business, consumer, mobility, fixed. When the 5G network is built, you have a fixed and mobile play that’s basically native to the deployment you’re doing.”

That means Verizon’s millimeter wave 5G network is designed to be shared by everyone and everything, including businesses, residential customers, cell phone users on the go, Internet of Things applications like smart meters and intelligent traffic systems, and more. But that network will not be everywhere Verizon or Verizon Wireless currently provides service.

“Our deployments of millimeter wave are focused on urban centers. It’s where the people are, where the consumption is,” Koeppe said.

Verizon faces significant costs building out its 5G wireless network in areas where it does not already offer FiOS fiber to the home service. Verizon’s 5G network is dependent on a fiber optic-fed network of small cells placed on top of utility and light poles at least every few city blocks. That means Verizon is most likely to get a reasonable return on its investment placing its 5G network in urban downtown areas and high wireless traffic suburban zones, such as around event venues, large shopping centers and entertainment districts. The company has chosen to deploy 5G in some residential areas, but only within large city limits. So far, it has generally steered clear of residential suburbs in favor of older gentrified city neighborhoods with plenty of closely-spaced multi-dwelling apartments, condos, and homes, as well as in urban centers with converted lofts or apartments.

Koeppe

Rural areas are definitely off Verizon’s list because the millimeter waves Verizon prefers to use do not travel very far, making it very expensive to deploy the technology to serve a relatively small number of customers.

Other carriers are not committing to large scale 5G deployments either.

At a debate held earlier today at Georgetown Law’s Institute for Technology Law & Policy, former FCC commissioner Mignon Clyburn, now a paid lobbyist for T-Mobile, warned that unless T-Mobile was allowed to merge with Sprint, its deployment of 5G will only happen in “very limited areas.”

Sprint has plans to introduce its own flavor of 5G, which won’t use millimeter wave frequencies, by June in nine U.S. cities. T-Mobile has talked about deploying 5G on existing large cell towers, which means one tower will serve many more customers than Verizon’s small cells. But with more customers sharing that bandwidth, the effective speed customers will see is likely to be only incrementally better than T-Mobile’s existing 4G LTE network. AT&T is initially moving in the same direction as T-Mobile, meaning many customers will be sharing the same bandwidth. That may explain why AT&T’s current 5G hotspot service plan also comes with a 15 GB data cap.

Verizon says its millimeter wave network will, by geography and design, limit the number of people sharing each small cell, making data caps unnecessary for its 5G fixed wireless home broadband replacement, which delivers download speeds of around 300 Mbps on average.

“We engineer the network to give the customer what they need when they need it, and the results speak for themselves,” Koeppe said.

Verizon is already selling its 5G service in limited areas for $50 a month to Verizon Wireless customers, $70 a month for non-customers. There are no data caps or speed throttles.

Based on the plans of all four major U.S. carriers, consumers should only expect scattered rollouts of 5G this year, and only in certain neighborhoods at first. It will take several years to build out the different iterations of 5G technology, with millimeter wave taking the longest to expand because of infrastructure and potential permitting issues.

Investors are not buying into the substantial hype surrounding the forthcoming 5G revolution and many remain unconvinced about the benefits of spending billions of investor dollars to deploy the next generation in wireless.

A survey by telecom analyst McKinsey & Co., picked up a clear drag on proposed spending, especially outside of North America, as carriers are finding investors reluctant about the business case for 5G technology.

The survey found 60% of wireless operators cited selling investors on the merits of 5G to be their greatest challenge. Only 25% were confident they could successfully prove a substantial return on investment for shareholders who typically want short-term results. Investors are demanding detailed evidence that 5G networks, the most costly requiring large fiber optic networks and neighborhood small cell antennas, will pay off with increased revenue and customer demand. Unlike earlier cellular standards which required incremental upgrades usually on existing cell towers, the fastest iteration of 5G will require providers to construct costly new fiber networks with a very large number of short-range antennas expected to be placed on top of utility or light poles.

Customer demand for 5G is anticipated to be low until new devices are introduced capable of connecting to it, and investors already recognize consumers are increasingly delaying device upgrades since the industry dropped two-year service contracts with device subsidies. Ongoing upgrades to existing 4G LTE networks may ultimately dampen demand even for less costly 5G networks that will be deployed on existing cell towers. McKinsey’s survey found less than 35% of respondents are planning quick launches of 5G on gigabit-speed capable millimeter wave spectrum, citing the high cost of deploying small cell networks.

“Until th[e business case emerges], most operators will tread cautiously, leveraging 5G for near-term objectives and waiting for a clearer view on the use cases’ economics to accelerate,” the McKinsey report concluded. “Given the expense required to prove those significant use cases, it could be an uncomfortably long wait. And for operators in countries that don’t see 5G as a matter of strategic and economic importance, there is a greater risk of falling behind.”

In early 2019, most network operators will focus on network planning and funding, with the first significant wave of launches expected in the U.S. coming later this year. Just over half of U.S. operators plan to have “large-scale 5G deployments” completed by late this year, and U.S. carriers are the most optimistic 5G can make good business sense, at least for some applications. Other U.S. carriers expect their networks to launch by the end of 2022. But in neither case are those launches expected to be widespread across the country. Competing with cable and phone company internet with fixed wireless service is also a non-priority for most operators.

“At least at the outset, the majority see enhanced mobile broadband and the Internet of Things (IoT), rather than fixed wireless access or mission-critical applications, as the most prevalent applications,” McKinsey’s survey found. Among early potential applications are smart utility meter connectivity, traffic sensing, and connected public infrastructure like lighting and traffic control signals. Giving consumers a way out of choosing between Verizon and Comcast for home internet service is not going to be an early priority for companies like AT&T. In fact, starting a price war is the last thing investors want to see.

“Although commercially in its infancy, 5G technology is ready, and in most markets its presence will be felt from 2020 on,” McKinsey’s report states. “Yet the fact that commercial models are not ready cannot be minimized; the business case is marginal, and the investments to enable new business models are not currently planned.”

McKinsey believes what will ultimately drive a gradual rollout of 5G technology is network congestion which can no longer be managed through existing traditional cell tower networks, known as “macro sites.”

“In rural and suburban areas, as well as along roadways, operators can handle increased traffic simply by densifying existing networks with macro sites,” McKinsey shared. “In many highly populated urban areas, by contrast, they’ll need to rely on small-cell solutions for two reasons: a higher concentration of traffic, as measured by traffic load per square kilometer, and the use of higher spectrum bands (greater than 3 gigahertz).”

But making the jump from the traditional large cell tower to a network of small cells scattered around neighborhoods will require a great deal of money. Operators will need to build fiber optic connectivity to each small cell, which can be managed either with a newly constructed fiber project or leasing existing fiber optic networks, presumably from cable operators which already have a significant fiber presence. In either scenario, rural areas will largely be left out, because all-important network traffic density is generally inadequate to support the business case for 5G, and cable operators are unlikely to have fiber networks available to lease in those areas.

Windstream is relying on the Federal Communications Commission’s Connect America Fund to double the areas where it will offer 100 Mbps broadband service, expected to reach 30% of the company’s 18-state local service area by the end of the first quarter of 2019.

“Windstream understands that premium internet speeds are critical to families and businesses in rural America, and we are systematically enhancing our network to meet that urgent demand,” said Jeff Small, president of consumer and small and medium-sized business services. “Network upgrades are expensive, especially in rural areas where there are relatively few customers, so Windstream is using a combination of its own capital and crucial support from the FCC’s Connect America Fund to make faster speeds more widely available. Without support from the Connect America Fund, many of these projects simply would not be economically feasible.”

Thomas told attendees at the Citi 2019 TMT West Conference Windstream’s legacy copper wire telephone network is not up to the job of handling the kinds of internet speeds more modern technologies can manage.

In urban and larger service areas, Windstream is most likely to deploy fiber to the home service in new housing developments and select gentrified neighborhoods where a business case exists to invest in fiber upgrades. The company also typically replaces its copper wireline infrastructure with fiber where road construction projects or damage forces the company to replace or relocate its lines. Suburban and more densely populated rural areas are likely to receive an upgraded version of Windstream’s DSL service that can manage up to 50 or 100 Mbps. In Windstream’s significant rural service area, the phone company is increasingly turning to fixed wireless technology, especially in flat midwestern states like Nebraska and Iowa where it plans to offer a combination of 3.5 GHz “CBRS” and 5G millimeter wave fixed wireless broadband capable of delivering up to 1,000 Mbps.

Windstream’s service area

“[We are deploying wireless internet] probably at a larger scale than a lot of the larger wireless companies,” Thomas said, especially in flatter areas where wireless signals go a long way.

Because most current broadband expansion fund programs require companies to commit to at least 25/3 Mbps service, simply expanding basic ADSL technology has proven inadequate to meet the government’s speed requirements. But wiring fiber to the home service to get faster speeds in rural areas does not meet the Return On Investment requirements Windstream’s shareholders demand. Windstream claims fixed wireless can solve both problems.

“You can get 100 Mbps out there very cost-effectively,” Thomas claimed. “You are really blowing away copper infrastructure and making it irrelevant because you’re embracing this 100 Mbps technology.”

As of early 2019, Windstream claims that 60% of its customers can get at least 25 Mbps service, 40% can receive at least 50 Mbps service. By the end of March, 30% will be able to receive 100 Mbps service.

A satisfied Windstream customer talks about his upgrade to 50/8 Mbps, which replaces his old 6 Mbps DSL service. (6:03)