Thursday, July 6, 2023

How Much FTTH Investment Will Prove Excessive?

Some observers might note that there have been periods of overinvestment in various connectivity sectors since 1995, and in data center capacity to a lesser extent. 


Subsea capacity and competitive local exchange carrier segments in the 1995 to 2001 period come to mind, as well as low earth orbit satellite systems in the early 2000s as well. 


Global Crossing, Level 3 Communications, 360Networks, NorthPoint Communications, and Winstar Communications were a few of the subsea or CLEC firms that went bankrupt. 


In 1998 Teledesic raised $9 billion in funding to launch a constellation of 288 satellites, but filed for bankruptcy in 2002. In 1999, Iridium launched a constellation of 66 satellites to provide global mobile satellite services and was bought out  by Globalstar in 2007.


In 2000, Orbcomm launched a constellation of 28 satellites to provide two-way data messaging services. The company survived as a small participant in the LEO business.


In the mobile segment, Metropcs, Nextel and VoiceStream declared bankruptcy about the same time. 


And though data center capacity has in the past briefly been overbuilt, demand has relatively quickly absorbed the supply. 


Data Center Market

Years

Degree of oversupply

Northern Virginia

2015-2017

20%-30%

Silicon Valley

2016-2018

15%-25%

Frankfurt

2017-2019

10%-20%

Singapore

2018-2020

5%-15%


Some might ask questions about the sustainability of many fiber-to-home business plans now underway in the U.S. market. In the past, firms have come to grief when they borrowed too much money, were overly optimistic about their sales and revenues and faced too much competition. 


Consider one scenario where overinvestment in FTTH proves troublesome. Keep in mind that investment to create facilities does not mean customer acceptance. Lumen, for example, sells FTTH to about 26 percent of locations passed. Verizon has peaked at just a bit over 40 percent, as have most other telcos with FTTH footprints. The typical pattern is lower take rates when service is launched, with higher terminal rates after three years of marketing in any single market.  


While revenue sources for some ISPs with extensive FTTH networks will include revenue from business customers and wholesale operations, many ISPs are looking at payback models anchored by home broadband services with monthly revenues between $50 to $80 a month. 


Whether you believe that is sustainable for an ISP with 80 percent or greater market share and other revenue sources, questions begin to grow as the terminal market share forecasts in competitive markets dip towards 20 percent or 30 percent. 


Though there is certain to be huge disagreement about such forecasts, one might argue that the home broadband market is moving towards a much-reduced role for fiber-to-home platforms and a heightened role for mobile and other wireless access technologies, as the number and types of connected devices grows, with greater reliance on mobile or untethered access, and increasing ability to leverage mobile network assets to support both mobile and fixed use cases. 


Some rival platform executives might be more optimistic about their chances of competing with FTTH. Most customers, for example, do not buy the fastest-available services but rather “good enough” services that cost less. Also, we should expect every platform to continue to increase provided speeds over time. 


Some observers might expect satellite access to remain at no more than 10 percent. And though FTTH should gain, so will mobile-only and fixed wireless. Perhaps most in the “things will change, but not drastically” camp believe that HFC share (cable operator) will decline as more customers buy FTTH instead. 


Platform

2023 Market Share

2030 Market Share

Fiber to Home

15%

20%

Hybrid Fiber Coax (HFC)

60%

50%

Fixed Wireless

5%

10%

Mobile-Only

15%

20%


Fiber supporters will argue that FTTH will do far better, eventually perhaps representing 40 percent share of market by 2030, as all fixed network ISPs shift to FTTH platforms, including cable operators. In such scenarios, perhaps FTTH holds 40 percent share (including share of telcos, cable and independent ISPs). HFC could drop to 30 percent while shares of the other platforms share the rest of the market. 


The real sensitivity in the model is how fast new FTTH lines can be added by 2030, compared to how fast cable operators can continue to upgrade HFC service, or themselves switch to FTTH. 


The argument for FTTH is that it is the only futureproof platform for internet access. And while that might prove correct. But it might be a longer transition than many now expect.


No comments:

Have LLMs Hit an Improvement Wall, or Not?

Some might argue it is way too early to worry about a slowdown in large language model performance improvement rates . But some already voic...