For South Korean FTTH, Definitions Matter

Sometimes definitions matter; sometimes they are less relevant. In the U.S. market, it does not much matter whether access media is fiber to the premises or hybrid fiber coax or wireless. The issue is how much bandwidth consumers can buy, and at what cost, irrespective of access media.

In the South Korean internet access market, one is confronted by apparently contradictory statistics: fiber to premises supply is quite high, but take rates for fiber to home services is relatively low. 

On one hand, fiber to the premises is said to reach 80 percent to 98 percent of locations. On the other hand, take rates are said to be as low as 10 percent, 38 percent or as high as 98 percent. 

The other issue is speed. Just because a connection uses optical fiber media does not mean the supplied services run at especially high speeds. Some telcos that installed FTTH systems back in the late 1990s were supplying service at 10 Mbps. In South Korea, in 2017, average FTTH speeds were as low as 29 Mbps. 

In 2019, according to Ookla, South Korea average internet access speeds were about 144 Mbps downstream. 

source: FTTH Council Asia Pacific

So something odd has to be explained: FTTH supply is quite high, and yet average speeds are perhaps lower than one might expect. That explanation probably has to rely on the high percentage of multiple dwelling units in South Korea. 

Apartments represent perhaps 52 percent to 60 percent of dwellings in South Korea, and arguably is as high as 70 percent in Seoul. And that is likely a key to understanding the data. By definition, an optical fiber connection to a high-rise building goes into the basement, with actual end user access over a copper connection. 

source: FTTH Council Asia Pacific

In a strict sense, one might argue that the actual end user connection therefore uses copper cables, not optical fiber.

Slow Revenue Growth Remains Key Industry Problem

Revenue growth continues to be the big connectivity industry issue, with key markets looking at one percent to two percent annual revenue increases, according to S&P Global. 

source: S&P Global

Microsoft Plans to be Carbon Negative

By 2030 Microsoft will be carbon negative, and by 2050 Microsoft will remove from the environment all the carbon the company has emitted either directly or by electrical consumption since it was founded in 1975, Microsoft says. 

Microsoft aims to  cut its carbon emissions by more than half by 2030, “both for our direct emissions and for our entire supply and value chain,” the firm says. “We will fund this in part by expanding our internal carbon fee, in place since 2012 and increased last year, to start charging not only our direct emissions, but those from our supply and value chains.”

It also is noteworthy that Microsoft intends to ground its efforts in “ongoing scientific advances and an accurate reliance on the basic but fundamental mathematical concepts involved.”

How Well is U.S. Fixed Network Business Doing?

In the U.S. market, there is a fairly clear bifurcation between fixed network connectivity providers doing relatively well, and those which are not doing so well, even if an argument can be made that the segment as a whole is challenged. 

Tier-one cable operators are generating more revenue and profits from their wireline operations than most telcos are. Large tier-one telcos are doing better than smaller telcos. But AT&T fixed network revenue is slowly declining. 

Verizon fixed network revenue also has been declining slowly. 

source: Verizon

The broader traditional telco market is doing less well, as revenue is dropping. 

S&P Global Ratings

Cable internet service revenue growth arguably exceeds telco growth, which is flat to negative. And cable revenue from fixed network operations is growing. “Overall, we expect top-line growth of about five percent with weighted average margins improving by about 50 bps to approach 39 percent, S&P Global estimates. 

S&P Global Ratings

“We believe cable companies can continue to increase high-margin broadband revenue for the next two years through a combination of subscriber growth and higher prices as subscribers demand faster internet speeds with rising data consumption,” says S&P Global. “We expect that the average number of broadband subscribers will increase by about four percent in 2020.”

S&P Global Ratings

WAN Costs Might be a Bigger Edge Computing Driver than Latency

By 2025, perhaps 75 percent of enterprise data will be generated outside enterprise data centers or cloud data centers, according to Cisco. 

While today only 10 percent of all data is handled at the edge, analysts expect in three years between 50 percent and 75 percent of all data to be produced and processed at the edge,” said Paul Morgan, Global Sales for Manufacturing, Automotive & IoT, at HP Enterprise (HPE).  “Gartner puts the figure at 75 percent.”

And even if latency is an issue for some edge applications, conrtainment of bandwidth costs also matters. If much of that data can be processed locally, wide area network bandwidth costs are lower.

Spectrum Per Customer Matters

One always has to take marketing claims with a grain of salt. So it is with U.S. mobile operator spectrum holdings.

AT&T recently has claimed a dramatic lead in low-band spectrum. This chart shows why AT&T makes the claim. 

AT&T has about 176 MHz worth of low-band and mid-band spectrum, compared to Verizon’s 117 Mhz, Sprint’s 212 Mhz, T-Mobile’s 11o MHz and Dish Network’s 92 MHz. 

Of course, Verizon and AT&T have the most subscribers, so bandwidth per subscriber is less than for Sprint, T-Mobile US or Dish Network. 

source: Wells Fargo Securities

The “spectrum per subscriber” picture is different, though, because one also has to factor in the network load. Operators with more customers "need" more spectrum. And since Verizon and AT&T have "most of the customers," that should affect spectrum available "per customer."

source: Mobile Experts

Spectrum holdings matter, of course. But subscriber loading also matters. Looked at on a bandwidth per subscriber basis, AT&T, Verizon and T-Mobile US are not far apart. Only Sprint has an unusually high amount of spectrum. Dish Network has not launched yet, and will be starting with modest network loading, so it should have relatively high spectrum per customer.

Verizon has 35 percent share of subscriptions. AT&T has 34 percent market share. So those two service providers have a combined 69 percent share of market. T-Mobile US has 17.5 percent share, while Sprint has about 12 percent share, according to Statista.

5G Subscription Forecasts Suggest Asia Will Lead

Market forecasts always are contingent on one’s assumptions, and that is no different for 5G subscription forecasts. It would take a brave leap or unusual definition to conclude anything other than that Asia will have the greatest number of 5G connections in the future, simply because Asia has the most people and the most mobile subscribers. 

Juniper Research anticipates that over 75 percent of global 5G connections will be in the Far East and China. This is due to the early launches in South Korea by all tier one operators, which was followed by significant launches of commercial 5G networks in China. 

GlobalData, on the other hand, suggests that Asia will account for about 65 percent of global 5G subscriptions by 2024 and about 44 percent of revenue. North America will account for 32 percent of revenue by 2024.

At the end of 2019, there were an estimated 4.5 million 5G users in the Far East, roughly 80 percent of the global total, nearly all of them in South Korea, Juniper Research estimates. 

But Juniper also predicts the United States and South Korea will be the fastest adopters of 5G, with 75 percent of all 5G subscribers attributable to these two countries by the end of 2020. That does not make sense to me, but that is what Juniper says. 

By 2025, Asia will still have more than 60 percent of 5G connections, Ericsson predicts. 

source: Ericsson