Thursday, August 30, 2018

In Indianapolis, Consumer Internet Access Market is About to Become Unstable

AT&T says it will launch mobile 5G in Indianapolis this year. AT&T previously had said Atlanta, Charlotte, Dallas, Oklahoma City, Raleigh and Waco would be early to get 5G as well. Keep in mind that AT&T already launched 5G Evolution in Indianapolis in 2017.

That means a peak theoretical wireless speed of at least 400 Mbps, for devices that can handle LTE-LAA.  Using LTE-LAA, the network has peak theoretical speeds for capable devices reaching up to 1 gigabit per second.

Some will note that Indianapolis also is a market where Verizon says it will launch 5G fixed wireless services that will compete with AT&T and Comcast fixed network internet access.

So here’s the intriguing new development. We are used to measuring internet access market share in silos: share of fixed network accounts and then share of mobile accounts. In Indianapolis and perhaps dozens of other cities, that will not be so easy, in the future.

Verizon will essentially use its 5G in fixed mode to compete for consumer and business fixed network access accounts. And now AT&T might find its own mobile 5G services will become a substitute product for its own fixed network services.

The amount of market share change among fixed and mobile; cable and telco providers might change quite significantly, forcing analysts to look at consumer segment  internet access market share in a much-ore comprehensive way.

How, for example, do we characterize “typical” or “average” internet access speed? In some cases, the mobile speeds will be higher than fixed network speeds.

Comcast in Indianapolis sells consumer internet access at speeds up to 350 Mbps, and offers a 2-Gbps service that likely will mostly be purchased by businesses.

So the issue now is how providers package their offers when it is conceivable that either fixed or mobile 5G offers speeds equivalent with the best Comcast has to offer.
    
To the extent that cable TV industry fortunes now rely on internet access revenues, and to the extent that new competition emerges from 5G fixed wireless, that growth engine is exposed.

“We see 5G fixed wireless broadband as the biggest existential threat to broadband providers (by far),” say equity analysts at Cowen. Assume, for example, that T-Mobile US actually throws significant effort at 5G fixed wireless, and gets anywhere near its goal of 10 million accounts by about 2024.

“That would be a large majority of the entire cable industry’s broadband adds over the next six years,” Cowen analysts say.

Verizon, for its part, is targeting 30 million or perhaps 35 million homes passed, in metro markets where it has lots of fiber (places where XO has big footprint, for example) and outside its legacy fixed network territory.

That suggests Verizon could wind up competing most with AT&T, Comcast and Charter. Making a further assumption that in its chosen markets it will be Comcast and Charter that have the most market share, and already have the greatest share of faster-speed accounts, that logicall suggests the cable competitors might be at more risk than AT&T or CenturyLink.

You might argue that the telcos generally sell lower-speed services, so their customers would be most at risk from a fixed wireless entry by Verizon or T-Mobile. But most of those customers already could have switched to cable providers, were internet access speed the big consideration.

What might be the case is that customers of slower-speed telco services think those services are good enough, when bundled with other telco services, to inhibit switching behavior.


Verizon has significant metro fiber in 17 of the top “NFL” U.S. markets where it is not the incumbent.

Assuming the cable competitor in those markets has 45 percent share, IComcast could have 2.1 million homes exposed to loss, about (eight percent of its base).

Charter might be at risk to share loss of perhaps 1.1 million homes, about 4.7 percent of its base.

Assuming a 25 percent Verizon take rate when it enters those new markets, perhaps two percent of Comcast customers and one percent of Charter customers could go to Verizon. That might be a manageable loss.

T-Mobile US might take a different tack. T -Mobile expects to become a viable in -home broadband alternative especially in rural areas. The carrier believes it will, by 2021, it provide data rates in excess of 100 Mbps to 66 percent of the U.S. population, reaching coverage of up to 90 percent by 2024.

T-Mobile US execs have suggested that a service offering 100 Mbps would be a competitive alternative for perhaps 19 percent  of the population. By 2024, as much as 35 percent to 45 percent of the U.S. population might be candidates for a mobile substitute product.

Looking at fixed wireless, T-Mobile US estimates it could provide service to 9.5 million households early on. By 2024 T -Mobile expects to be able to reach 52 million rural residents with a fixed wireless solution.


Some who decry the lack of competition in the U.S. internet access space might be surprised to see growing competition, at scale, when 5G arrives. Google Fiber once was thought to be a catalyst for such competition. Now it appears Verizon and possibly T-Mobile US will assume that role.

Sometimes it takes scale to disrupt a market. In Indianapolis, that might well be that the actor is Verizon. But AT&T's response might also be disruptive to Comcast.

We are entering a period of greater instability in the consumer internet access business than we have seen for quite some time (decades, in fact).

Wednesday, August 29, 2018

Wild Variance in Estimates of U.S. Fixed Network Broadband Access

Assumptions matter. Consider estimates of internet access using fixed networks of all types. Some have estimated the total number of homes (accounts) as low as 46 million, in 2015. Other estimates (2018 figures) put the number of U.S. fixed network internet access accounts at a minimum of 97 million.

Growth rates for fixed network internet access simply are not high enough to get to 97 million from 46 million in several years. So those high and low estimates have to be making different assumptions.

There are many possible reasons for the sharp divergence. Since the definition of “broadband” changes from time to time, one plausible explanation is that the lower estimates are based on the latest and “fastest” definitions. But estimates of broadband adoption seem to vary wildly.

One estimate from the World Bank in 2015 suggested broadband adoption was only about 30 percent in the U.S. market, which few would consider close to reality. An estimate based on U.S. Census Bureau data from about the same time pegged home use of the internet at 75 percent.

Another plausible explanation is that some researchers use a “homes that buy broadband” figure, which speaks to buy rates, not the “I can buy, but choose not to buy” figure, which represents the “I could buy if I wanted to” number.

Yet other estimates might simply count the number of either “homes passed” by networks offering internet access service or the number of buyers of internet access at any speed.

All those choices matter.




The point is that one has to be very careful about assumptions about how many people, homes, or businesses have access to internet access or broadband; how many people have access; what speeds they can purchase; what speeds they do purchase; or what percentage of potential buyers have supplier choice or how many are choosing to use platforms other than a cabled network.

One fact seems clear enough: the number of homes and businesses able to buy a reasonably-fast internet access service continues to inch upwards. In the 5G era, the number of locations able to buy much-faster services should take a step function upwards, with order of magnitude speed increases (10 times) for many U.S. households.

In the meantime, incremental progress is being made.

Some 700,000 U.S. rural homes and small businesses will gain access to high-speed internet access for the first time through the Federal Communications Commission’s Connect America Fund Phase II auction, auction results released today show, and more than half of those 713,176 locations will have service available with download speeds of at least 100 megabits per second.

Connecting 700,000 homes in a nation with perhaps 138.3 million housing units, that might not sound like much. But the 250,000 most-isolated rural locations account for about half the estimated cost of connecting seven million U.S. homes in rural, hard to reach areas.

Of course, the most-isolated areas likely always will be among the best candidates for satellite or other wireless access platforms. Perhaps three million rural locations already buy satellite internet access.

That “cost to reach” logically implies that the great bulk of new investment under this program will support extension of service to new locations that are not the most-isolated, as that provides the greatest efficiency. In other words, service providers logically will connect the new locations in some proximity to current networks, rather than taking on the most-difficult, most-expensive areas .

According to plan, 53 percent of all homes and businesses served with support from the auction will have broadband available with download speeds of at least 100 megabits per second. Some 19 percent will have gigabit service available. And 711,389 locations—all but 0.25 percent—will have at least 25 Mbps service.

It appears a great many of the providers are ISPs using fixed wireless.  

Providers must build out to 40 percent of the assigned homes and businesses in a state within three years of becoming authorized to receive support. Buildout must increase by 20 percent in each subsequent year, until complete buildout is reach at the end of the sixth year.

The auction allocated $1.488 billion in support to be distributed over the next 10 years to expand rural broadband service in unserved areas in 45 states. A total of 103 providers will get funds.

Tuesday, August 28, 2018

6G?

It is too early to speculate about what a 6G mobile network might look like, what features it might add, over 5G, or what applications or use cases might emerge on such a network. But the ways the computing industry has changed might offer some insight.

Even if performance tends to improve with time, we eventually reached a point where the raw increases in processor speed, of the cost of memory, became less important in shaping the business than where, how and why computing was done. In other words, cloud computing and  mobility, spoken interfaces and so forth arguably are more important than processing speed.

So when 5G has been deployed ubiquitously, and consumers have access to download speeds as high as 10 Gbps, one-millisecond latency, an order of magnitude (10 times) to two orders of magnitude (100 times) change in the cost of supplying internet bandwidth, with network slicing and edge computing, it really is a bit difficult now to figure out how much additional effort needs to be put into raw speed, improved latency or cost per bit.

Instead, the focus should shift, as did computing, to where, how and why communications gets done, and then optimizing 6G to support those changes. It just seems unlikely, at this point, that raw increases in speed, latency or other core network performance features will have the same impact as once was the case from deploying faster networks.

Cost Per Bit Will Drive Most of the Early 5G Use Cases

In the early going, an order of magnitude (10 times) to two orders of magnitude (100 times) change in the cost of supplying internet bandwidth might be the key early value of 5G networks. In other words, vastly-lower cost per bit is going to drive early revenue model changes.

That might be true even if latency performance and speed also change by an order of magnitude.

The reason is that the step change in cost per bit supports the existing business case (supplying more bandwidth) to mobile phones and other devices, as well as creating the foundation for at least a couple new revenue sources, without requiring the creation of other brand-new services or applications that harness the ultra-low latency or faster speeds.

Where demand for bandwidth is highest (the 10 percent to 15 percent of cell locations where traffic demand is greatest), 5G can be spot deployed to improve user experience.

In the category of “new use cases,” the earliest addition will be the use of 5G in fixed wireless mode to attack the fixed network internet access customer base.

The extent to which a pricing premium for “faster speeds” allows a 5G price premium, how much and for how long, remains to be seen, but there will be some likely impact there.


Monday, August 27, 2018

Less Competition in U.S. Mobile Market?

Even if one is of the opinion that merger rules are more relaxed now than under the prior presidential administration, some quantitative tests always are used by the Department of Justice when evaluating proposed mergers, and immediate past precedent might be a key issue.

The proposed merger of AT&T and T-Mobile in 2011 was scuttled because the Hirschman Herfindahl Index (“HHI”) would have increased by 700 points in an industry already classified by the HHI and highly concentrated. The HHI is a standard tool used by regulators globally and will be used again as Sprint and T-Mobile US propose to merge.

It is not incorrect to read the DoJ’s objections to that earlier merger as opposition to any merger between two of the four nationwide providers of mobile wireless services: AT&T, Verizon, T-Mobile and Sprint.

The proposed merger of the third- and fourth- largest U.S. mobile companies would increase the HHI by over 400 points, meaning market concentration, in an industry already deemed highly concentrated (if less concentrated than some other markets), would increase.

“Like the failed AT&T and T-Mobile merger a few years ago, the proposed merger between Sprint and T-Mobile will be presumed anticompetitive and thus heavily scrutinized by antitrust authorities,” says George Ford, Phoenix Center for Advanced Legal and Economic Public Policy Studies chief economist.

To be sure, executives from T-Mobile US and Sprint say the additional scale provided by the merger will allow the bigger firm to continue to be aggressive in attacking prices, or will allow the firm to invest in 5G at a faster rate.

But I have yet to find an equity analyst who believes that will happen. Quite the reverse is expected, in fact: less price competition in the U.S. mobile market.

Telstra Launching Own IoT Tracking Services

Telstra is launching its own location tracking service using Internet of Things (IoT) sensors for Australian consumers, small businesses and enterprise customers. Customers will be able to attach Telstra Locator Tags to their valuables and use the Telstra Locator App to help find them if they go missing.

Tags include a lightweight and long-lasting Bluetooth tag for keys and purses and a rechargeable Wi-Fi tag ideal for pets, bikes and bags. Early next year Telstra will introduce an LTE tag, designed for high-value assets such as vehicles and machinery.


The Telstra Device Locator feature for Telstra’s 24x7 App allows customers locate compatible Telstra mobile phones and SIM-enabled tablets on their Telstra account, including the ability to send tailored messages to lost devices and play sounds to locate devices in hard to see places.

Telstra is also working with a number of large organisations to trial its new Track and Monitor solution for enterprise customers, which will be available from October 2018.

Enterprise customers will be able to purchase Track and Monitor to enable low-cost, large-volume asset tracking, whether across multiple warehouses or retail sites,  or while in transit.


The announcement follows Telstra’s deployment of Cat-M1 and Narrowband IoT technology on its mobile network over the past 12 months, the company says.

Friday, August 24, 2018

What's the Revenue Upside for 5G Network Slicing?

Why 5G network slicing--the ability to create multiple logical networks to run on top of a shared physical network infrastructure--matters to enterprises and service providers is a logical question.

Other existing technologies supply some of the value of virtualized networks supporting network slices. Differentiated Services (DiffServ) supplies quality of service. Virtual Private Networks already create logical networks using  IP tunneling.


Network Function Virtualization (NFV) already is coming to mobile core networks.

So what is the business value of network slicing? New revenue sources are conceivable, for starters.

Mobile operators can partition their physical network resources into multiple logical slices and lease those slices out to interested parties. Think of how mobile virtual network operators now purchase wholesale capacity from network suppliers, and then consider how an MVNO might use a network slice to source its capacity using a network slice, instead of buying capacity the traditional way.

Also new: the possibility of selling a range of capabilities ranging from a fully private network to use of core network, transport or portions of access networks with customized features.


That could allow an MVNO to control its core network, not simply use radio and other generic resources obtained wholesale from a network services provider. An MVNO might add other resources (computing resources, distributed or centralized; plus custom or other apps).

An electrical utility might configure its own network slice for connectivity for sensors, meters, and controllers that do not require much bandwidth at all, and operate mostly in fixed mode.

As with the long-sought advantage of networks supporting bandwidth and provisioning on demand, network slicing could offer an easier way to provision short-term and episodic networks, such as temporary networks set up for events. A temporary network created to support  a concert or music festival might prefer a network slice that is optimized to support streaming.


Video surveillance firms might want their slices configured to support cameras.

Other potential slice buyers could include big app or transaction platforms that could assemble complete wholesale mobile networks by buying network slices.

At least in a few cases, some upstart mobile network providers might elect to take a wholesale-only approach, offering their customers virtual networks possibly optimized differently for various MVNO tenants.

There are corresponding new revenue models as well. Multi-end-user slices might created for customers who have similar requirements, such as high-bandwidth, low-latency or any other parameter the network can provider (quality of service, availability, resilience, security, identity management,     geography, low device power consumption). That could happen across some industry verticals, for example (health, manufacturing, content delivery)

Temporary networks provided by slices or single-enterprise or single-organization networks also can be created and sold.

Mobile operators also might choose to use network slicing for internal use only, as a wholesale platform, as a building block for vertical market services.

Thursday, August 23, 2018

5G Virtualizes the Whole Network

In a statement about whether Australian service providers may buy infrastructure products, Australian lawmakers make some other points about changes to be wrought by 5G. Others have noted that 5G is not simply an air interface, but a change in how logical functions in the core network flow through the access network.

“5G requires a change in way the network operates compared to previous mobile generations,” say Australian Senator Mitch Fifield, Minister for Communications and the Arts, and Scott Morrison, member of Parliament and Acting Minister for Home Affairs. Put simply,  “the distinction between the core and the edge will disappear over time,” at least in terms of security, control and other logical network functions.

‘Where previous mobile networks featured clear functional divisions between the core and the edge, 5G is designed so that sensitive functions currently performed in the physically and logically separated core will gradually move closer to the edge of the network,” they note.

In other words, the network is virtualized.



Wednesday, August 22, 2018

Investment Risk in U.K. and Elsewhere in 5G Era

An agency of the U.K. government wants 15 million premises connected by optical fiber by 2025, with coverage across all parts of the country by 2033, as well as 5G coverage of “the majority of the population” by 2027.

As always, those goals set by the U.K. Department for Digital, Culture, Media & Sport are a statement of direction: actual investments and timetables will likely vary. But it also might be fair to say that the amount of investment risk larger than in the past, for several reasons.

One key issue is that U.K. service provider revenue has decreased 1.7 percent a year since 2012. That raises the usual questions about additional investment in an industry with shrinking revenue.


The report sees some amount of facilities-based competition.  “At least a third (with the potential to be substantially higher) of U.K. premises are likely to be able to support three or more competing gigabit-capable networks,” the report predicts. “Up to half (or lower if there are more three network areas) of premises are likely to be in areas that can support competition between two gigabit-capable networks.”

That poses the “normal risk” of stranded assets (service providers will not earn revenue from every home or business the network is built to support.

A convergence of fixed and mobile infrastructure also is foreseen. “In the longer-term, the government expects to see a more converged telecoms sector,” as fixed networks and 5G are complementary.

That means suppliers about to leverage both mobile and fixed assets could gain advantages.

Also, for the first time at scale, 5G could be a replacement for fixed networks. “The technology synergies between 5G and fixed networks are likely to create strategic advantages for those operators that have interests in both,” the report says. And one potential advantage is the ability to deploy commercial gigabit networks at lower cost than using fiber-to-home networks.

The industry has not, to this point, faced a scenario where mobile networks literally are a full and functional substitute for fixed networks, which have had the advantages of speed and much-lower cost per bit.

The point is that capital investment decisions carry higher risk, as competition could take new forms, at scale, and alter payback periods, account potential and profit margin.

But risk exists on the revenue side as well. Among the biggest unknowns is the amount and extent of revenue upside, with much of the benefit to come in the form of capacity gains, as mobile operators deploy 5G first where they face the greatest data demand.

Other new services will develop over time, it is assumed and hoped. “There is some uncertainty over the business models for 5G,” the report suggests. That will tend to lead investors to be careful about how fast, and how much, to invest in new 5G facilities. To limit risk, it will make sense to deploy 5G first  in dense urban areas where capacity is an existing concern.

Such moves will better match investment to revenue upside.

Also, 5G might “also open up opportunities for new players to enter the U.K. market,” by which the report authors mean new neutral host infrastructure providers at venues, rural or outdoor small cell coverage. That could again affect business model assumptions about investment cost.

Is Sora an "iPhone Moment?"

Sora is OpenAI’s new cutting-edge and possibly disruptive AI model that can generate realistic videos based on textual descriptions.  Perhap...