The Next Generation of the Internet is Coming

You might argue the internet has evolved since its inception. Originally a narrowband tool for researchers, it now is a broadband tool used widely by most consumers and businesses. Early on in the development of the World Wide Web, people gained the ability to publish, says Charles Fan, Cheetah Mobile CTO. Then, with the emergence of search, we gained the ability to find the world's information, he says. The problem is that you have to know what you are looking for, to find it. That makes search less useful.

In the coming wave, content relevant to a consumer will be found and then "pushed" to each user. We already see glimmers of that in the high use of social apps, where people now find "information" useful and relevant to them. Also, "news" is redefined less as what is happening in the broader world, and more what is happening with your friends, family and social circles.

To a large extent, that means we are using an algorithm that essentially assumes "what is interesting to your friends is interesting to you." That is correct, up to a point. The next wave will involve use of artificial intelligence, coupled with big data stores, to actually predict what you like.

In the next generation of the internet, machine learning will be better than knowing your social graph, as a way of connecting you with things you are interested in. That AI-driven model might also lead to creation of huge new business models to replace existing and older models (advertising, e-commerce or peer trading mechanisms like Uber), says Fan.

That reliance on big data stores might have huge implications. On one hand, the algorithms will have huge amounts of new data to work with. On the other hand, algorithms will be commoditized, democratized or “made somewhat obsolete.” In a world where everyone has access to good algorithms, value will shift to access to huge data stores.

“Who has the better quality of data wins,” Fan says. “How do you get better at collecting data?” Fan asks. “Big data ownership is going to be the oil of the modern age.” But that will require artificial intelligence or machine learning.

For most of us, artificial intelligence (AI) has been a science project for the past few decades: interesting and provocative, but not something that actually affects the businesses most of us deal with on a daily or even annual basis.

There are reasons to believe that is changing. Charles Fan, Cheetah Mobile CTO, said it might seem odd for a mobile app and tools company such as Cheetah Mobile to be seriously evaluating AI. Actually, it turns out to be most practical, as Cheetah Mobile launches new applications in the news aggregation area.

Eventually, the ability to personalize and then predict what a particular person might like will require AI to mine and then predict and deliver “suggested items” to individual people.

Your social profile helps, but only so much. Content providers or advertisers can assume you are somewhat like your “friends” in terms of interests. But only up to a point is that correct. Each individual actually is quite different, at a more-granular level. But it will take AI to rapidly process all the data used to assemble a highly-personalized set of content and then match people with highly-targeted offers and ads.

Incentive Auction a "Massive Disappointment?"

Among the conclusions one might draw from the Federal Communications Commission’s 600 MHz two-stage “incentive auction” are that low-band mobile spectrum, in the U.S. market, is deemed to be worth less than always has been argued.

In terms of revenue raised by the new way of auctioning spectrum, some might call the outcome a massive disappointment.

In January 2016, FCC Chairman Tom Wheeler claimed the auction would be the “world's largest spectrum auction that has ever taken place.” Not so, as it turns out. In fact, the auction likely will clear far less (less than half) of the $44.9 billion raised by auctioning AWS-3 spectrum.

By way of comparison, the AWS-3 spectrum (in the 1.7 GHz and 2.1 GHz ranges) raised $44.9 billion for 65 MHz of total spectrum.

The 600-MHz spectrum will raise perhaps $18 billion for 84 MHz of spectrum. The FCC originally had expected to raise as much as $60 billion in proceeds.

This was in sharp contrast to the FCC’s initial expectation of collecting at least $60 billion from Incentive Auction. Spectrum owners had expected as much as $86 billion in sales.

Arguably, AT&T and Verizon believe they have other ways to satisfy any “coverage” capacity they might require in the future, while the “reserve spectrum” conditions that barred both from bidding on much of the spectrum (about a third) might have contributed to the lack of enthusiasm as well.

Some argued that the "reserve spectrum" feature, as always, produces distortions in spectrum markets. Among those distortions are price and demand impacts.

On the other hand, some bidders such as T-Mobile US will pay less for their spectrum, and likely will get some spectrum, simply because competitors were forbidden to do so.

There are a huge number of ways mobile operators will be increasing capacity in coming years. There is an astounding amount of new millimeter wave spectrum coming that dwarfs all current authorizations of communications spectrum.

Offload of demand to Wi-Fi also will help. So will small cell networks operators will build to deploy the new millimeter wave assets. As much as seven gigaHertz of new unlicensed spectrum also will be released by the FCC, allowing use without payment of license fees.

At the same time, better radios also will help boost capacity, using any specific set of frequencies. And some potential bidders, such as cable companies, also can buy existing companies and thereby acquire their spectrum assets.

We will have to wait to see whether other large auctions of new spectrum (India and Egypt  also have found far less demand than anticipated) continue to show lower prices.

But it seems markets are rational. With vast increases in supply now possible, and more supply coming, prices should fall.

Where Should India Focus Internet Access Capital Investment?

Some might say, in terms of fixed network internet access, across South Asia. Others might say that is less important than achieving robust mobile access at higher speeds, since that is the way most people are going to use internet access services.

Perhaps paradoxically, India and others should not deploy excessive capital on fixed facilities, instead focusing on mobile access at higher speeds, one might argue. The best way to get everyone connected is to use the mobile and wireless platforms that can be deployed and used faster, at lower cost and lower investment levels, right now.

Infrastructure is a major barrier, among several, for Indian internet access ubiquity, according to McKinsey and Company. In fact, according to McKinsey, infrastructure is the single biggest barrier to widespread internet access.

source: TRAI

source: TRAI

"No material impact" if Strong Net Neutrality Rules Go Away: Netflix

“Weakening of US net neutrality laws, should that occur, is unlikely to materially affect our domestic margins or service quality,” says Reed Hastings, Netflix CEO. Oddly enough, cable TV companies said the same about common carrier regulations: they did not prevent those firms from investing in faster access services. Telcos clearly reduced capital investment.

But the incentives and behavior arguably fall into a couple of buckets. Netflix believes it clearly benefits from rules that mandate “best effort only” access, in large part because that prevents creation of quality of service (QoS) mechanisms that could raise Netflix costs of doing business.

Telcos believe they clearly are harmed, because they cannot create new QoS tiers of service that could create new revenue models (business services, with QoS assurances, have higher gross revenue and profit margins, for example).

Cable companies, positioned to gain market share because they could quickly improve speeds, were in position to do so even if no QoS were possible, simply because their networks could rapidly and affordably be boosted into multi-hundreds of megabits per second ranges, plus gigabit speeds. Despite the rules, cable was able to justify new investments.

So there is no easy way to assess the impact of the net neutrality rules in some uniform sense. Netflix arguably benefited, but no longer needs the rules. Cable companies also benefited, despite the rules. Telcos, already losing the market share battle, arguably were dissuaded from investing.

The point that might be made is that network neutrality has been a blunt instrument with different impact on different participants in the market. Going forward, and without arguing for imposing new rules on app providers in advance of demonstrable problems, a more nuanced approach could work.

Competition in access markets, for example, would be enhanced if telcos were convinced they could do something--especially in the quality area--to compete more effectively with cable companies, and would invest to do so.

The new perspective that will at least get a hearing is that net neutrality or other rules designed to protect consumers from abuse of market power--the sorts of policies normally the domain of the Federal Trade Commission--might need to be symmetrical. That is to say, they perhaps should apply to edge providers (app providers) and access providers alike.

Netflix continues to favor” strong net neutrality” rules (no QoS). That arguably will not happen, which is why Netflix reassures its investors that it can thrive even without the rules.

Artificial Intelligence is Becoming Quite Practical

For most of us, artificial intelligence (AI) has been a science project for the past few decades: interesting and provocative, but not something that actually affects the businesses most of us deal with on a daily orFor most of us, artificial intelligence (AI) has been a science project for the past few decades: interesting and provocative, but not something that actually affects the businesses most of us deal with on a daily or even annual basis.

There are reasons to believe that is changing. Charles Fan, Cheetah Mobile CTO, said it might seem odd for a mobile app and tools company such as Cheetah Mobile to be seriously evaluating AI. Actually, it turns out to be most practical, as Cheetah Mobile launches new applications in the news aggregation area.

Eventually, the ability to personalize and then predict what a particular person might like will require AI to mine and then predict and deliver “suggested items” to individual people.

Your social profile helps, but only so much. Content providers or advertisers can assume you are somewhat like your “friends” in terms of interests. But only up to a point is that correct. Each individual actually is quite different, at a more-granular level. But it will take AI to rapidly process all the data used to assemble a highly-personalized set of content and then match people with highly-targeted offers and ads. even annual basis.

There are reasons to believe that is changing. Charles Fan, Cheetah Mobile CTO, said it might seem odd for a mobile app and tools company such as Cheetah Mobile to be seriously evaluating AI. Actually, it turns out to be most practical, as Cheetah Mobile launches new applications in the news aggregation area.

Eventually, the ability to personalize and then predict what a particular person might like will require AI to mine and then predict and deliver “suggested items” to individual people.

Your social profile helps, but only so much. Content providers or advertisers can assume you are somewhat like your “friends” in terms of interests. But only up to a point is that correct. Each individual actually is quite different, at a more-granular level. But it will take AI to rapidly process all the data used to assemble a highly-personalized set of content and then match people with highly-targeted offers and ads.

Verizon Buying Comcast or Charter--Absent Huge Divestitures--Would Not Pass Horizontal Antitrust Review

Rumors about big mega-mergers get floated from time to time for all sorts of reasons, sometimes to test the waters or encourage deal making. That seems to be the case for rumors about Verizon buying Comcast, or Verizon buying Charter Communications. Either combination would clearly run into major antitrust issues using the standard screens used by the Justice Department.

So such rumors might be intended to gauge just how far antitrust rules can be bent. But barring major divestiture of assets, it seems virtually impossible that antitrust officials would have an easy time with such huge horizontal mergers.

Some might try to argue that Verizon fixed network triple play services and Comcast or Charter triple play services do not represent horizontal concentration. Or Verizon might contemplate divesting its whole fixed network operation, though it is hard to see who the logical buyers might be.

Some might like to point out that AT&T bought Tele-Communications Inc., the biggest US. cable TV company, but that was totally different. AT&T (not SBC) had zero consumer fixed network assets. So buying TCI did not bring AT&T more than about 30 percent access to U.S. homes.

SBC, now that it has acquired the former AT&T, did so only after AT&T sold off those former TCI assets to Comcast, allowing Comcast to become the biggest U.S. cable TV company.

As Comcast and Charter now are the largest U.S. ISPs, and the largest U.S. cable TV companies, adding Verizon’s fixed network customer base and either cable company’s households passed would easily surpass the 30 percent maximum reach that historically has been the size limit for any local access provider.

Verizon could try and divest enough Comcast or Charter assets to stay under, or at, the 30-percent reach of U.S. households, keeping all its own fixed network assets.

Still, all those would be complex divestitures, with other complications. It isn’t so clear how much Verizon would want to operate hybrid fiber coax, all copper and all-fiber access networks, with all the differences in outside plant operations that would represent.

Nor is it so clear that Verizon would want to get bigger in fixed networks, since it has built its business on mobile services. Likewise, it is not so clear Verizon wants to be as big in media as owning Comcast would entail.

Charter’s footprint, meanwhile, is more rural than Comcast’s footprint, something Verizon definitely would not prefer.

To be sure, assets could be divested. They would almost certainly have to be, on the fixed network side of the business. Then there are all the cultural issues, union and non-union workforces to blend.

It seems unlikely Verizon would be looking strictly for a vertical merger, as AT&T proposed with Time Warner. Talk of Verizon using the HFC network for small cell and other backhaul makes sense, but the issue is the 30-percent homes passed test.

It all seems too complex, too likely to draw antitrust rejection, to be a likely outcome.

Reliance Jio Sets Interim Mobile Data Price of US$1.50 for 4 GB Per Day

Though we might still have to wait a while to find out what Reliance Jio do about “every day” pricing of mobile data after the six-month free trial it has been running, Jio now has launched another retail offer to replace the free data offer, creating a new mobile data plan sold for a nominal cost for the three-month period after the free data offers end at the end of March 2017.

The new offer,  valid till June 30, 2017,  will charge a fee of about Rs 100 (about US$1.50) for data, while voice will be free.

Depending on when customers signed up, the monthly usage allowance can be as much as 4 GB per day, or as little as 1 GB per day.