Will MEC Revenues Ever Surpass Capex to Create it?

One recurring issue with forecasts of multi-access edge computing is that it is easier to make predictions about cost than revenue and infrastructure investments than customer demand. Beyond that, MEC almost necessarily involves some sharing of revenue among ecosystem participants. 

To determine whether the connectivity provider business model for MEC actually works, one has to delineate cost (capital investment as well as operating costs) as well as revenue. And some of us still do not clearly see how the business plan works out, at least where it comes to retail revenue.

It is a bit easier to envision how edge computing works as a necessary part of the network infrastructure to support advanced mobile networks, which are, by definition these days, distributed. In other words, MEC might be just as important as telco operating infrastructure as a possible source of retail revenue from customers.

Research and Markets has forecast 2028 MEC revenue at about $23 billion. That mostly includes infrastructure sold by hardware and software suppliers to create MEC capabilities. 

Service revenue from actual “edge computing” is included, but one still must make assumptions about the receipt of such revenue by ecosystem partners, such as a hyperscale computing as a service supplier and its access network partners. 

But one matter is clear enough: most “MEC” revenue is generated by the sale of hardware and software to create the capability, not revenues earned by the supplier of the actual “computing” service. 

source: Research and Markets 

It matters greatly “who” is supplying the edge computing “as a service” and who is providing other necessary infrastructure. It also matters “what” is defined as “edge computing revenue and services.” In addition to the “as a service” revenue, there are recurring revenue contributions from real estate (colocation) and connectivity. 

Right now, it is nearly impossible to accurately predict the magnitude of “edge computing” revenues. How does Amazon Web Services state revenue for a customer account that uses both cloud and edge resources, for example? 

Does a mobile operator report additional connectivity revenue supporting internet of things use cases as “connectivity” revenue in its business customer segment, or as “edge computing” revenues?

Looking only at private networks, there are similar issues. How do we separate “edge” from other “computing infrastructure” spending? As always, attributions matter. What is “private edge;”what is use of “public edge?”

 It is far easier--if not “easy”--to estimate what it might cost to create MEC capabilities. 

Private network multi-access edge computing investments will amount to about $6 billion by 2030, ABI Research now estimates.  

Juniper Research predicts that annual investments by telecommunications operators on multi-access edge computing infrastructure will reach $11.6 billion in 2027, up from $5.4 billion in 2022; a CAGR of 16.7 percent.

Total “spending” on  MEC facilities will be higher. Juniper Research estimates investment of nearly $9 billion in 2022 growing to nearly $23 billion in 2027. 

By 2027 mobile operators will have built out more than 3.4 million MEC nodes, up from less than one million by the end of 2022. 

Juniper forecasts that over 1.6 billion mobile users will have access to services underpinned by MEC nodes by 2027, up from only 390 million in 2022. 

Of course, that is a prediction about capital investment, not revenue. And revenue also is a complicated matter where it comes to edge computing. Edge computing spending can represent purchases of user or network devices; software capabilities or computing as a service or 5G access to support edge computing. 

So revenue can accrue to a number of participants in the edge computing ecosystem: device retailers, network infrastructure providers, software suppliers or connectivity service providers. So when researchers talk about MEC revenue or investments, one has to separate shares within the ecosystem. 

source: Grand View Research 

Some estimates have total MEC revenues exceeding $25 billion by perhaps 2027 and close to $70 billion by 2032.  Other estimates suggest annual revenue of close to $17 billion by 2027.  

But those forecasts virtually always lump together revenues earned by hardware, software and services suppliers: infrastructure and platform plus computing as a service revenues. And computing as a service revenues will likely be dominated by hyperscalers, not mobile operators. 

Connectivity providers will profit from real estate support and some increase in connectivity revenues, but relatively rarely from the actual “edge computing as a service” revenues. 

For example, assume 2021 MEC revenues of $1.6 billion globally; a cumulative average growth rate of 33 percent per year; services share of 30 percent; telco share of service revenue at 10 percent. 

Multi-access Edge Computing Forecast
Year	2021	2022	2023	2024	2025	2026	2027	2028
Revenue $B	1.6	2.1	2.8	3.8	5.0	6.7	8.9	11.8
Services Share	0.3	0.6	0.8	1.1	1.5	2.0	2.7	3.5
Growth Rate	0.33
Telco Share	0.1
Telco Revenue	0.2	0.2	0.3	0.4	0.5	0.7	0.9	1.2
source: IP Carrier

The actual MEC revenue from MEC is quite small by 2028. In fact, too small to measure. Of course, all forecasts are about assumptions. 

One can assume higher or lower growth rates; different amounts of connectivity provider participation in the services business; different telco shares of the actual “computing as a service” revenue stream; greater or lesser contributions from mobile connectivity revenue from MEC. 

The point is that actual MEC revenues earned by mobile operators or other connectivity providers might actually be quite low. So value earned from all those infrastructure investments would have to come in other ways.

Higher subscription rates; higher profit margins; lower churn; higher average revenue per account are some of the ways MEC could provide a return on invested capital. Some service providers might actually provide the “computing as a service” function as well, in which case MEC revenues could be two to three times higher. 

But many observers are likely to be disappointed by the actual direct revenue MEC creates for a connectivity provider.  

So revenue can accrue to a number of participants in the edge computing ecosystem: device retailers, network infrastructure providers, software suppliers or connectivity service providers. So when researchers talk about MEC revenue or investments, one has to separate shares within the ecosystem. 

How Much Incremental New Revenue for Mobile Operators from Edge Computing?

As important as they might be, incremental new revenue streams from 5G, internet of things, private networks and edge computing might possibly be far smaller than many assume. For starters, the bulk of 5G revenues will continue to come from business and consumer "phone" subscriptions.

New connectiions supporting IoT, private networks or edge computing will take time to develop, and might not create as much revenue as some project. IoT average revenue per device, for example, is more than an order of magnitude lower--and often nearly two orders of magnitude lower--than for a consumer mobile phone connectiion. So volume matters.

5G private networks will not necessarily be "managed" services supplied by mobile operators. Enterprises will build their own networks, as they build their own Wi-Fi and other local networks.

And multi-access edge computing business models are generally shifting towards partnerships with hyperscale cloud computing suppliers. That means most of the "edge computing service revenue" will be earned by the hyperscalers, not mobile operators.

But value and revenue for mobile operators are expected from all those areas. The issue is how much it will cost to participate, and the magnitude of revenue streams. Consider the investment side.

Juniper Research predicts that annual investments by telecommunications operators on multi-access edge computing infrastructure will reach $11.6 billion in 2027, up from $5.4 billion in 2022; a CAGR of 16.7 percent.

Total “spending” on MEC facilities will be higher. Juniper Research estimates investment of nearly $9 billion in 2022 growing to nearly $23 billion in 2027.

By 2027 mobile operators will have built out more than 3.4 million MEC nodes, up from less than one million by the end of 2022.

Juniper forecasts that over 1.6 billion mobile users will have access to services underpinned by MEC nodes by 2027, up from only 390 million in 2022.

Of course, that is a prediction about capital investment, not revenue. And revenue also is a complicated matter where it comes to edge computing. Edge computing spending can represent purchases of user or network devices; software capabilities or computing as a service or 5G access to support edge computing.

So revenue can accrue to a number of participants in the edge computing ecosystem: device retailers, network infrastructure providers, software suppliers or connectivity service providers. So when researchers talk about MEC revenue or investments, one has to separate shares within the ecosystem.



source: Grand View Research

Some estimates have total MEC revenues exceeding $25 billion by perhaps 2027 and close to $70 billion by 2032. Other estimates suggest annual revenue of close to $17 billion by 2027.

But those forecasts virtually always lump together revenues earned by hardware, software and services suppliers: infrastructure and platform plus computing as a service revenues. And computing as a service revenues will likely be dominated by hyperscalers, not mobile operators.

Connectivity providers will profit from real estate support and some increase in connectivity revenues, but relatively rarely from the actual “edge computing as a service” revenues.

For example, assume 2021 MEC revenues of $1.6 billion globally; a cumulative average growth rate of 33 percent per year; services share of 30 percent; telco share of service revenue at 10 percent.

The actual MEC revenue from MEC is quite small by 2028. In fact, too small to measure. Of course, all forecasts are about assumptions.

Multi-access Edge Computing Forecast $Billions
Year	2021	2022	2023	2024	2025	2026	2027	2028
Revenue $B	1.6	2.1	2.8	3.8	5.0	6.7	8.9	11.8
Services Share	0.3	0.6	0.8	1.1	1.5	2.0	2.7	3.5
Growth Rate	0.33
Telco Share	0.1
Telco Revenue	0.2	0.2	0.3	0.4	0.5	0.7	0.9	1.2
source: IP Carrier

One can assume higher or lower growth rates; different amounts of connectivity provider participation in the services business; different telco shares of the actual “computing as a service” revenue stream; greater or lesser contributions from mobile connectivity revenue from MEC.

The point is that actual MEC revenues earned by mobile operators or other connectivity providers might actually be quite low. So value earned from all those infrastructure investments would have to come in other ways.

Higher subscription rates; higher profit margins; lower churn; higher average revenue per account are some of the ways MEC could provide a return on invested capital. Some service providers might actually provide the “computing as a service” function as well, in which case MEC revenues could be two to three times higher.

But many observers are likely to be disappointed by the actual direct revenue MEC creates for a connectivity provider.

Throwing Stones When Living in Glass Houses

It is hard to find anybody who argues the Australian National Broadband Network has been a success, after a decade of turmoil. On the other hand, broadband upgrades have presented a difficult business case in many large countries with relatively low density and serious competition. As easy as it might be to criticize apparent failure, the difficulty is not to be underestimated. 

 

Since wholesale networks seem to have worked elsewhere, albeit in smaller countries, it is hard to argue that the wholesale model itself is the issue.

To be sure, high construction costs are a real issue in a country as large as Australia, with so many low-density areas. But many would say the unexpectedly high costs have other causes.

The Australian Communications Commission plans to revamp wholesale pricing. But ACCC believes the NBN--even if changes are made--will not be profitable until 2040. The new plan hopes to recoup accumulated investment losses by that point. 

Some might argue the better course is to amidst failure and allow the equivalent of a bankruptcy. There seems no political appetite for doing so, despite the advantages. For more than 20 years, private connectivity firms that have amassed unworkable levels of debt have been allowed to fail. 

There is much pain for lenders and shareholders when this happens, but successful restructurings that eliminate the debt problem often allow firms to create business models that have a chance of succeeding. 

In other cases firms simply find they are unable to continue, and assets are sold. Many forget that the company once known as AT&T was itself acquired by SBC, which rebranded itself after the acquisition. 

That demise was, in turn, fueled by high debt AT&T took on in an effort to rapidly create a national access strategy based largely on the use of cable TV physical plant. 

The strategy failed, in part because the debt burden was too high; in part because the cost of upgrading the cable TV plant proved too daunting in the timeframe AT&T required. 

AT&T’s big move into cable TV in the mid-1990s came at a time when the long distance provider was seeking a way to reenter the local access business with its own facilities. The thinking at that time was that a largely one-way cable TV plant could be upgraded to become full communications facilities, supporting home broadband and voice. 

Given that development by virtually all cable TV companies in North America and Europe, the thinking was sound. But timing and huge capital investment costs were issues.

On June 24, 1998, AT&T acquired Tele-Communications Inc. for $48 billion, marking a reentry by AT&T into the local access business it had been barred from since 1984. 

Beside TCI, at that point the largest U.S. cable company, AT&T also bought MediaOne, the cable asset holding of US West. 

US West had in 1994 purchased Wometco and GTC cable operations and then Continental Cablevision, creating MediaOne Group. 

US West also participated in a joint venture agreement with Time Warner Cable to form Time Warner Communications (later known as TW Telecom).

In 1993 MediaOne purchased a 26 percent stake in Time Warner's entertainment operations including Warner Bros. and HBO. Those assets also wound up at AT&T. 

AT&T also purchased Teleport Communications Group, a $500-million-a-year local business phone company, for $13.3 billion; MetroNet, a Canadian phone system, for $7 billion; and the IBM Global Network, which carries data traffic, for $5 billion. 

AT&T also signed a joint venture with Time Warner Cable ( to carry phone calls over the entertainment conglomerate's cable TV systems, and with British Telecom to serve multinationals overseas. 

Nor was AT&T the only telco to consider the cable TV strategy. Since 1994, major telcos had been discussing--and making--acquisitions of cable TV assets. In 1992 TCI came close to selling itself to Bell Atlantic, a forerunner of Verizon. Cox Cable in 1994 discussed merging with Southwestern Bell, though the deal was not consummated. 

US West made its first cable TV acquisitions in 1994 as well. In 1995 several major U.S. telcos made acquisitions of fixed wireless companies, hoping to leverage that platform to enter the video entertainment business. Bell Atlantic Corp. and NYNEX Corp. invested $100 million in CAI Wireless Systems.

Pacific Telesis paid $175 million for Cross Country Wireless Cable in Riverside, Calif.; and another $160 to $175 million for MMDS channels owned by Transworld Holdings and Videotron in California and other locations. 

AT&T also made its first investment in DirecTV in 1996, owned and spun off Liberty Media. 

But the debt burden was too high and AT&T reversed course in 2004 and sold most of those assets. AT&T Broadband (the former TCI and US West Broadband assets) were sold to Comcast, making that firm the biggest U.S. cable TV company. 

By 2005 AT&T itself was acquired by SBC Communications, which promptly rebranded itself AT&T. 

The point is that broadband upgrade strategies that might have seemed reasonable at the time often turned out to be unworkable in the time frame required, with the hoped-for costs. Excessive debt almost always was among the chief problems. 

As it turned out, there often was no quick, affordable alternative for telcos looking to create broadband networks that were ubiquitous, capital efficient and timely. 

That “get there fast” mentality remains, as the cost and time to completely revamp the fixed network plant remains an issue. That is why 5G fixed wireless now looms so large. Some believe 5G and future untethered networks will provide further complications to NBN profitability. 

No matter how one looks at the business case, fixed network broadband remains expensive, time consuming, with a difficult business case. It never is surprising that firms look for shortcuts.

The whole argument made by cable operators, in fact, has been that hybrid fiber coax provides a faster, cheaper way of getting to gigabit broadband on a mass scale, compared to fiber to the home. 

Multi-gigabit speeds are coming, so a tough business case will not be getting much easier.

Actually, Consumers Might be Relatively Resistent to Connectivity Service Provider Switches

Nobody would be surprised if told that, in general, consumers will switch products and suppliers for a better price or better value.

But consumers also intuitively understand that some savings are more important than others. In other words, there might be limits to the amount of effort most consumers will put into comparing connectivity service offers. 

Put another way, there is only so much a consumer can save by aggressive shopping for connectivity services. So it appears many do not bother. Some studies suggest that typical churn rates now are fairly low in many markets, for many services. Contracts and high prices, as well as intensity of usage, are some drivers of higher churn, as you might guess. 

source: Researchgate 

While Australians are happy to switch for a better price or customer experience, almost 50 per cent of surveyed respondents admit to doing either no research or a basic level of research before choosing their internet provider (47 per cent), according to Commonwealth Bank research.  

Switching barriers seem to exist. Some 60 percent say they worry competitive offers will be the same, or perhaps worse. That is probably much more true for experienced consumers who have had multiple suppliers in the past. 

Not to be discounted: there is a learning curve with any new provider. Once a customer has become familiar, switching barriers increase, which is why 42 percent of those surveyed say they are comfortable with their existing providers. 

Once a customer has figured out that a current provider supplies the expected value at a reasonable price, with acceptable customer service, account longevity actually is a good predictor of future “low churn” risk. 

Issues with service quality, in contrast, are a very good predictor of high churn risk. 

source 

The point is that although consumers are expected to prefer “saving money,” the incentives to research alternate providers are low, relative to search effort. Most consumers likely perceive a zone of reasonableness where value and price are concerned.

Most consumers rightly perceive that offers from competitors most often tend to be equivalent in many respects. 

“The common perception is that changing Internet providers is more hassle than it’s worth,” More CEO Andrew Branson says. 

Most consumers likely see switching costs in the form of new gear that has to be purchased or leased, set-up charges, possible contract requirements, bundling with other services that might also have to be changed, and uncertainty about intangibles such as customer service ease, signal quality and consistency.

At any given point in time, perhaps 90 percent of consumers are probably satisfied enough that unhappiness is not driving them to consider switching connectivity providers. In the U.S. mobile market, for example, fewer than eight percent of consumers say they are “likely” to switch service providers at any given point in time. A smaller percentage actually do so. 

source: Bain 

Switching behaviors for products that have little switching cost are robust, one might argue.  

For fixed services, much switching behavior is related simply to a household move from one area to another that also requires switching service providers.   

Does 5G Violate Network Neutrality Principles? COAI Head Says it Does

For every public purpose there are corresponding private interests. The same might be said of connectivity provider regulation. That is the only explanation for a mobile operator association arguing a key 5G capability inherently violates network neutrality principles.

It is so odd an argument that some other agenda must be at work.

Is network slicing--a new feature of the fully-deployed 5G network--a threat to “network neutrality?” Possibly, says SP Kochhar, Cellular Operators Association of India general director. At first glance, it is an odd position for a mobile industry group to take. 

"In essence, the main network will be like economy class, and ones derived out of slices with different parameters can be business class or first class,” says Kochhar. 

Built into the 5G standard is the ability of the core 5G network to create virtual private networks that can vary parameters such as latency and bandwidth prioritization. The industry has argued that such features allow creation of customized networks that are essentially “tuned” for use cases that are latency-dependent or bandwidth availability dependent. 

Industry proponents have argued that this creates new revenue potential for mobile operators. So what are we to make of COAI virtually arguing that network slicing violates network neutrality principles?

One has to work backwards. Since network neutrality prohibits “quality of service” mechanisms for consumer customers, Kochhar is virtually arguing that the 5G network core network should be prohibited from offering network-slicing-based services. 

What conceivable benefit is seen? 5G is possible without building out the full 5G core network: 5G end user services can be delivered over the 4G evolved packet core network. 

So perhaps COAI believes a bar on network slicing would mean India mobile operators could introduce 5G using the 4G core network, which would delay capital expenditures for a time. 

Also, mobile operators believe the ability of enterprises to acquire their own spectrum to set up private 5G networks is a dire threat to public 5G. They believe as much as 30 percent to 40 percent of enterprise mobility revenues are at stake. That seems an exaggeration, though it is conceivable that an enterprise 5G network could reduce demand for public 5G resources when users are “at work.” 

True “mobility” needs would not change because an enterprise 5G network exists. In essence, a private 5G network would allow organizations to offload public network traffic to the private network in the same way that they already can offload public network traffic to Wi-Fi. 

And most mobile operators consider that an advantage, not a problem. 

So it seems the invocation of the network neutrality “problem” is part of an effort to delay 5G core network requirements.

The Department of Telecommunications (DoT) defines net neutrality as the concept of non-discrimination of internet traffic by intermediate networks on any criteria. 

"The network should be neutral to all the information being transmitted through it. All communication passing through a network should be treated equally, independent of its content, application, service, device, sender or recipient address," DoT rules say.

Network neutrality means different things to different people, and is applied to different instances in different countries. The basic idea is that internet service providers (and not other entities) should provide nondiscriminatory consumer access to lawful internet content regardless of its source or destination. 

Generally speaking, internet access services sold to businesses are not covered. But many regulatory entities have added concepts. Some regulators hold that net neutrality also means ISPs cannot offer a “free tier” of service or allow sponsors to defray the cost of access to their services. 

But net neutrality is virtually nowhere a constraint on business customers. And network slicing is virtually never going to be a “consumer” service. It will be purchased by an enterprise or other organization, in the same way that content delivery network services are purchased by enterprises, not individuals and consumers. 

COAI has to understand that, so ithe comments about network slicing cannot be taken at face value. There is some other agenda.