By 2020, Most Software Products will Use Artificial Intelligence

By 2020, almost every new software product and service will incorporate artificial intelligence features, Gartner predicts.

Though it remains unclear how much business impact might eventually be derived, some early adopters already find AI contributes.

Amazon uses robotics to automate “picking and packing” activities in its warehouses, McKinsey notes. The “click to ship” cycle time, which ranged from 60 to 75 minutes with humans, fell to 15 minutes after applying robotics, while inventory capacity increased by 50 percent. Operating costs fell an estimated 20 percent, McKinsey argues.

Netflix uses AI to personalize recommendations. Netflix found that customers, on average, give up 90 seconds after searching for a movie. By improving search results, Netflix projects that they have avoided canceled subscriptions that would reduce its revenue by $1 billion annually.

Baidu and Google spent between $20 billion to $30 billion on AI in 2016, McKinsey says.

Healthcare, financial services, and professional services are seeing the greatest increase in their profit margins as a result of AI adoption, McKinsey argues.

The McKinsey Global Institute Study on Artificial Intelligence, The Next Digital Frontier also estimates that total annual external investment in AI at between $8 billion to $12 billion in 2016, with machine learning attracting nearly 60 percent of that investment.

Robotics and speech recognition are two of the most popular investment areas.

source: McKinsey

source: McKinsey

An Industry You Will Not Recognize, Within a Decade

You might not recognize the telecom industry within a decade.

The industry is likely to earn less revenue than at present. The only issue is how much less will be earned. So the industry will contract.

There might be 85 percent fewer telecom companies in business. There might be only five global carriers. Revenue growth might not be lead by new subscribers or even any of the access services (voice, messaging, data).

Services sold to humans might not drive revenue growth, either. And where growth has been driven by consumers, growth might in the next phase be driven by enterprises.

The changes will come fast.

No matter how you look at it, “eras” in the telecom industry are coming faster, and ending just as fast.

Consider voice: the era of traditional voice lasted more than a century. The “VoIP” era arguably lasted 25 years. The era we now are in will not even be characterized by “voice.”

Going forward, the internet, mobility and over-the-top apps are the ways to understand where we are. But even that will change.

The recent decade has been dominated by mobile revenues--subscription growth in developing markets and mobile internet access in developed markets. In the 5G era, the industry should not even be lead by mobile data.

The reason is simple: in the digital era, about every decade, the growth drivers have changed. In the 2G era, subscriber growth was the driver. In the 3G era, messaging, email access and mobile web drove incremental growth.

In the 4G era, mobile web and apps, internet access and video consumption have been key. In the 5G era, it is expected that services for non-human internet of things apps will drive incremental revenue growth.

Another way to characterize industry history is to note that fixed networks once represented the whole industry. Today, fixed line revenue and subscribers are fractions of mobile revenue and subscribers.

Industry structure also has passed through eras, from monopoly to mobile duopoly to competition. Many analysts now predict an era of vast consolidation will follow, within the next decade. How vast?

How about a reduction of global service providers from 800 to about 100, over the next seven years? That is what Bell Labs now predicts.

All those fast-coming, momentous developments are why the Pacific Telecommunications Council has created a new “Industry Transformation Boot Camp” from its existing Spectrum Futures and PTC Academy programs.

The week-long boot camp will provide industry professionals with a concentrated look at what changes are coming, why they are coming, what key business challenges exist and what strategic options firms face.

source: e-strategyblog

source: Karel

source: Quartz

Edge Computing a Mobile Operator Opportunity?

Many telcos have argued that owning data center assets would be a good way to complement connectivity services. Some have found the synergies less than truly compelling. The issue now is whether mobile edge computing will be different.

AT&T hopes so, thinking that its existing real estate portfolio (central offices, macro towers and small cells will provide locations for edge computing.

So AT&T will outfit those facilities with high-end graphics processing chips and other general purpose computers, to create new edge computing facilities.

AT&T believes it could someday embed these systems in everyday items like traffic lights and other infrastructure as well.

The point is that edge computing could provide a way for telcos to leverage existing assets to create new roles in the computing fabric and ecosystem that some might argue have eluded them in the centralized computing era.

Will Most IoT Connections Use Fixed Network Access (Wi-Fi and other)?

In the past, long-range, low-power networks have had leading share of market (55 percent) for machine-to-machine applications, with short-range solutions such as Wi-Fi having perhaps 35 percent share, with mobile networks having about 10-percent share.

How that might change as internet of things use cases proliferate is one issue. Maravedis, for example, believes that mobile platforms ultimately will have the largest share. Others believe low power wide area networks will do better than that, connecting perhaps 16 percent of IoT devices, while mobile networks connect some seven percent of devices.

That suggests most IoT devices will use other methods, such as Wi-Fi. That matters, in part, because it means most IoT connections might not increase the amount of revenue earned by access providers in a direct way.

source: Maravedis

source: Analysys Mason  

Short-range connection technologies will continue to represent the overwhelming percentage of internet of things connections through 2022, Ericsson predicts.

source: Ericsson

As Big as it Seems, Mobile Industry is Shrinking

Many observers, and many other parts of the internet ecosystem, believe and act as though the "telecom" or "mobile" industry is a powerful, dangerous gatekeeper. Ironically, those views come at a time when the mobile industry arguably has past its peak, in terms of industry revenues, and faces huge challenges that likely will lead to most suppliers leaving the market (being acquired, for the most part).

Some will see such consolidation as a sign of growing power. In actuality, such moves will highlight industry stress.

We now are five years past “peak revenue” for the global mobile business, some would argue.

Virtually all analysts who follow revenue and profit trends in the global telecom business will agree that, despite growth in some regions, at a global level revenue and revenue growth have become key issues. Indeed, revenue is the paramount industry issue, heading into the 5G era.

The signs are quite evident. Operator average revenue per user is declining, globally.

Competition is one reason. Over the top product substitution is another source of pressure. Regulatory intervention such as caps on retail and wholesale voice and data charges and on mobile termination rates also are an issue.

source: Juniper Research

There are other issues, though. As always seems to be the case, later users tend to spend less than earlier adopters. That puts pressure on overall ARPU.

The decline in certain regions is exaggerated because of the depreciation of local currency bundles against the dollar (eg in Latin America and Europe) as well.

Using 2015 as a  baseline, Juniper argues that while subscribers are expected to increase by 1.24 times their 2015 value by 2022, over the same period global ARPU is expected to fall to less than 75 percent of its 2015 value.  

Total revenue--not just ARPU--also is declining. Using 2012 (the year revenues peaked) as the baseline, by 2016 alone revenues have fallen to just 88 percent of peak values.

Also, mobile  operator revenues from international mobile roaming are expected to experience an 11% decline in 2017  in key markets including Europe, North America and Asia, Juniper argues.

Mobile roaming annual revenues, worth an estimated $54 billion in 2016, will decline to $48 billion in 2017.

We are 5 Years Past "Peak Mobile"

We now are five years past “peak revenue” for the global mobile business, some would argue.

Virtually all analysts who follow revenue and profit trends in the global telecom business will agree that, despite growth in some regions, at a global level revenue and revenue growth have become key issues. Indeed, revenue is the paramount industry issue, heading into the 5G era.

The signs are quite evident. Operator average revenue per user is declining, globally.

Competition is one reason. Over the top product substitution is another source of pressure. Regulatory intervention such as caps on retail and wholesale voice and data charges and on mobile termination rates also are an issue.

source: Juniper Research

There are other issues, though. As always seems to be the case, later users tend to spend less than earlier adopters. That puts pressure on overall ARPU.

The decline in certain regions is exaggerated because of the depreciation of local currency bundles against the dollar (in Latin America and Europe) as well.

Using 2015 as a  baseline, Juniper argues that while subscribers are expected to increase by 1.24 times their 2015 value by 2022, over the same period global ARPU is expected to fall to less than 75 percent of its 2015 value.  

Total revenue--not just ARPU--also is declining. Using 2012 (the year revenues peaked) as the baseline, by 2016 alone revenues have fallen to just 88 percent of peak values.

Also, mobile  operator revenues from international mobile roaming are expected to experience an 11% decline in 2017  in key markets including Europe, North America and Asia, Juniper argues.

Mobile roaming annual revenues, worth an estimated $54 billion in 2016, will decline to $48 billion in 2017.

The point is that we now are past the peak of the human-driven mobile services product life cycle.

How Big a Gamble are 5G Networks? Big.

Just how big a gamble are 5G networks? Big, according to Juniper Research. “Unlike 4G, there is no discernable use case that will encourage operators to roll out 5G networks,” Juniper says.

Beyond that, in the early going, most of the 5G revenue might well come from customers replacing 4G accounts with 5G, with some revenue lift, but no net increase in subscriptions.

So great is the risk that private investors might not make the move to 5G without subsidies. “Juniper anticipates that increased investment from governmental bodies will be needed to encourage the development of these networks, with the exception of North America.”

Juniper Research forecasts that total operator billed 5G revenues will rise to $269 billion by 2025, from $850 million in 2019, the anticipated first year of services. That represents a 161 percent compound annual growth rate over the first seven years of 5G services.

What that portends, in terms of new service revenues, is unclear, as most of that revenue will come from 4G customers switching to 5G, with no automatic net gain in revenue or accounts.

Juniper “anticipates that the majority of acquired 5G connections will be users upgrading from 4G connections.”  For that reason, Juniper does not expect an increase in the number of the active subscriber information modules (SIMs) from the introduction of 5G networks.

Juniper  predicts that 66 percent of all 5G operator-billed revenues will come from North America the Far East and China by 2025.

At least at first, Juniper believes mobile operators will charge a premium over 4G for 5G services.

Also, early adopters will tend to be those customers who spend the most, each month.

So in  the early years, the actual numbers of connected subscribers will be very low, Juniper suggests.

Those projections, it can be argued, miss the impact of services sold to non-human devices, which many believe represents most of the opportunity for incremental revenue, beyond new opportunities for fixed wireless services.