Unless a broadband provider can deliver the same kind of profitability earned by U.S. cable operators, don’t expect significant private investment in broadband expansion even if the company can easily turn a profit.

That was the argument brought to a House hearing on funding broadband infrastructure expansion by Craig Moffett, a Wall Street analyst at Moffett Nathanson.

“Infrastructure deployment requires the expectation of a healthy return on capital,” Moffett told the House Communications and Technology Subcommittee in a hearing this afternoon. “That should be taken as a given, but all too often, in my experience, the issue of return on capital is either ignored or misunderstood in policy forums. It is not a matter of whether a business is or isn’t profitable, it is instead a matter of whether it is sufficiently profitable to warrant the high levels of capital investment required for the deployment of infrastructure.”

Moffett pointed to the massive profits earned by cable operators Comcast, Time Warner Cable, Charter and Cablevision, all of which earned returns well in excess of their cost of capital, ranging from 13-33 percent. Moffett argued Wall Street has come to expect those kinds of returns, and investors will take a hard look at companies deploying new expensive networks against those that have largely paid back much of the capital costs incurred when their networks were built decades ago.

Moffett continued to criticize the broadband expansion being undertaken by large incumbent telephone companies that he claims does not earn attractive returns for their wireline businesses, even as they have introduced new services like faster broadband and television.

“For example, a decade after first undertaking their FiOS fiber-to-the-home buildout to 18 million homes, Verizon has not yet come close to earning a return in excess of their cost of capital,” said Moffett. “In 2014 their aggregate wired telecommunications business earned a paltry 1.2% return, against a cost of capital of roughly 5%. For the non-financial types in the room, that’s the equivalent of borrowing money at 5% interest in order to earn interest of 1%. That’s a good way to go bankrupt.”

Moffett was also critical of AT&T’s planned expansion of gigabit fiber broadband.

“AT&T has committed to the FCC to make fiber available to a total of 11.7 million locations in their footprint in order to make their acquisition of DirecTV more palatable to policy-makers, but it is hard to be optimistic that they will do much better this time around,” Moffett argued.

Moffett believes competition is bad for the profitable broadband business.

Moffett

“The broader take-away here is that the returns to be had from overbuilding – that is, being the second or third broadband provider in a given market – are generally poor,” Moffett said. “Let that sink in for a moment. Stated simply, it means that market forces are unlikely to yield a competitive broadband market. Neither, by the way, does wireless appear to offer the promise of imminent competition for incumbent broadband providers. Wireless networks simply aren’t engineered for the kind of sustained throughput required for a wired-broadband-replacement service.”

As a result, investors prefer that the broadband marketplace remain a monopoly or duopoly to guarantee the kinds of healthy returns they have earned for years, especially from the cable stocks Moffett has always favored in reports to his clients. Additional competition drives prices down, reducing profits, which in turn discourages investors who have high expectations their money will make them a lot more money.

Moffett’s arguments are largely based on broadband being a for-profit private enterprise, not a public infrastructure effort. But it does explain why there is a willingness to compete in large cities where network construction costs are lower and rural communities remain relatively unserved. As with electrification 100 years ago, investor-owned utilities were willing to wire large communities while ignoring rural farms and communities. Only after electricity was deemed a necessary utility did alternative means of funding, including member-owned co-ops and community-owned utilities finish electrifying areas private capital ignored.

Moffett’s guide to better broadband is based entirely on profitability — delivering enough profits and other returns to attract investors that will look elsewhere if costs become too high. Community-owned broadband avoids this dilemma by advocating for break-even or modestly profitable networks that focus on service, not investor-attractive profits.

Several members of Congress commented Moffett’s vision of broadband was discouraging, even depressing, because it seemed to be locked in a for-profit, private sector model that had few answers to offer for communities left behind. Moffett even warned against oversight and regulation of incumbent cable and phone companies, claiming it would further drive away private investment.

But broadband customers, Moffett admitted, will still pay the price for investor expectations.

“As everyone understands, the cable video business is facing unprecedented pressure,” Moffett testified. “Cord cutting has been talked about for years but is finally starting to show up in a meaningful way in the numbers. And soaring programming costs are eating away at video profit margins. From a cable operator’s perspective, the video business and the broadband business are opposite sides of the same coin. It is, after all, all one infrastructure. Pressure on the video profit pool will therefore naturally trigger a pricing response in broadband, where cable operators will have greater pricing leverage.”

Moffett said the kinds of rate hikes consumers used to pay for cable television now increasingly transferred to broadband customers is nothing nefarious. To keep investors happy, the kind of returns once earned from cable television will now have be delivered on the backs of broadband customers if Congress expects cable companies to continue upgrading and expanding their networks.

“All else being equal, that will mean that even new builds of broadband will become increasingly economically challenged and therefore will become less and less likely,” said Moffett. “Or they will simply have to sharply raise broadband prices.”

Moffett’s comments do come with some baggage, however. His clients pay for his advice and Moffett has been a long-time supporter of cable industry stocks. He has been a strong and natural advocate for a cable industry that faces only token opposition. He has browbeaten executives to start broadband usage caps and usage-based billing to further boost broadband profits, slammed telephone company competition in the cable business as financially reckless and unwarranted, and dismissed Google Fiber as a project designed to help Google’s public policy aims more than earn the search giant profits from the broadband business.

But Moffett has also been wrong in the past, particularly with respect to cord-cutting which he used to downplay as an urban legend and on the ease cable companies would be able to acquire and merge with each other.

Beyond all that, Moffett and his clients have a proverbial dog in the fight. After years of pumping cable stocks, suggestions that more competition for the cable industry is a good thing would simply be bad for business.

The magic of fiber

Ever wonder why Comcast, one of the strongest defenders of classic coaxial-based cable technology, is suddenly getting on board the fiber-to-the-home bandwagon? New research suggests if they don’t, their market share could fall to 40% or less if a serious fiber competitor arrives.

“There’s some sort of magic associated with fiber,” John Caezza, president of Arris’s Access Technologies division, told Multichannel News. “Everyone thinks it’s better than [cable technology].”

The risks to the cable industry are clear: be prepared to upgrade or face customer losses.

Craig Moffett of Moffett Nathanson has never been a cheerleader for fiber to the home service. In 2008, Moffett vilified Verizon for its investment in a major fiber upgrade we know today as FiOS to replace its aging copper infrastructure, complaining it was too expensive and was overkill for most residential customers. He was more tolerant of AT&T’s less-costly fiber to the neighborhood approach, dubbed U-verse, that still used traditional telephone lines to deliver service into the home. Because U-verse did not need AT&T to replace wiring at each customer location, the cost savings were considerable. But the cost-capability compromise left AT&T with a less robust platform, with broadband speeds initially limited to a maximum of around 24Mbps.

While phone companies like AT&T and Verizon were saddled with the enormous cost of tearing out decades-old obsolete phone wiring to varying degrees, the cable industry seemed well positioned with a mature, yet still recent hybrid fiber-coaxial (HFC) platform that was upgraded in the 1990s in many cities. While still partly reliant on the same RG-6 and RG-11 coaxial cable used since the first days of cable television, cable companies also invested in fiber optics to bring services from distant headends to each town, removing some of the copper from their networks without the huge expense of bringing fiber all the way to customer homes.

For Moffett, it was the cable industry that had the network with room to grow without spending huge amounts of capital on upgrades. He has touted cable stocks ever since.

Moffett

What worries Moffett now isn’t Google, Frontier, CenturyLink, or even Verizon. He’s concerned about AT&T.

As part of its commitment to win approval of its merger with DirecTV, AT&T promised regulators in June it would expand AT&T U-verse with GigaPower — AT&T’s gigabit fiber to the home upgrade — to at least 11.7 million homes, nine million more than it has ever promised before. Comcast has a 32% overlap with AT&T U-verse, compared to Time Warner Cable (26%), Charter Communications (32%), Bright House Networks (25%) and Cox Communications (25%). Comcast had promised faster broadband with the advent of DOCSIS 3.1 beginning as early as next year. But the company isn’t willing to wait around to watch AT&T and others steal its speed-craving customers. This spring, it promised 2Gbps Gigabit Pro fiber to the home service to customers living within 1/3rd of a mile of the nearest Comcast fiber line.

Some in the cable industry complain Google’s huge marketing operation has saddled cable broadband with a bad rap — ‘it’s yesterday’s news, with Google Fiber representing the future.’ The marketing war has been largely won by Google, they say, leaving consumers convinced fiber is the better and more reliable technology, and they need it more than the cable company.

Cable’s defense is to consider some marketing changes of its own — including the idea of dropping the name “cable” from the business altogether, because it implies older technology. But despite any name change, most cable companies will continue to rely on HFC infrastructure for at least several more years, despite claims they are bringing their own middle mile fiber networks closer to customers than ever. Cable operators now serve an average of 400 homes from each cable node. Some cable companies like Comcast plan to cut the number of customers sharing a node to around 100-125 homes, which means fewer customers will share the same broadband connection. But in the end, that will make cable comparable at best to a fiber to the neighborhood network, still hampered to some degree by the presence of legacy coaxial copper cable. The industry believes most consumers will never see the limitations, and for those that do, a limited fiber buildout with a steep installation fee may keep costs (and demand) down to those who need the fastest possible speeds and are willing to pay to get them.

That philosophy may still cost cable companies customers if a fiber competitor doesn’t have to compromise speed and performance and can afford to charge less.

The top 10 U.S. cable companies currently account for 60% of the residential broadband market and 86% of all broadband net additions in the first quarter of 2015, says Leichtman Research Group.

Moffett predicts cable broadband will only capture 40% of share in markets where it faces a fiber to the home competitor (Google, EPB, Greenlight, Verizon FiOS), 55% in markets served by a fiber to the neighborhood competitor (U-verse, Prism), and 60% where the competition only sells DSL (most Frontier, Windstream service areas). Nationwide, AT&T’s newest gigabit fiber commitment could cost the cable industry 2.4% of the whole residential broadband market, Moffett said.

Phil McKinney, president and CEO of CableLabs, believes DOCSIS 3.1 — the next standard for cable broadband — can easily stand toe to toe with fiber to the home providers.

McKinney

“I think it [HFC] has tremendous life, and we are going to be riding it all day long,” Werner said. DOCSIS 3.1 “is definitely going to be our go-to animal. Due to ubiquity, we can go out and virtually serve all of our [customers] very quickly.”

Cable companies claim their speed increases reach all of their customers in a given area at the same time without playing games with “fiberhoods” or waiting for incremental service upgrades common with Google Fiber or AT&T’s U-verse. Customers, the industry says, also appreciate DOCSIS upgrades bring no service disruption and nobody has to come to the home to install or upgrade service.

“The cable industry has more fiber in the ground than each fiber provider in the world,” McKinney argues. “If you look at total fiber strand miles, there’s more fiber under management and under control of the [cable] operators than anybody else combined.”

That may be true, but Moffett thinks it is only natural shareholders may eventually punish the stocks of cable operators that will face competition from AT&T’s U-verse with GigaPower. There is precedent. Cablevision serves customers in New York, Connecticut, and New Jersey and faces fierce competition from Verizon FiOS in most of its service areas. That competition has been brutal, occasionally made worse in periodic price wars. What may be protecting cable stocks so far is the fact AT&T competition will only affect, at most, 32% of the impacted cable operators’ service areas.

AT&T’s gigabit network has also proved itself to be more press release than performance, with very limited availability in the cities where it claims to be available. Verizon FiOS, in contrast, is widely available in most of Cablevision’s service area.

Still, Comcast is hoping it can hang on to premium customers who demand the very fastest speeds and performance with targeted fiber.

“Gigabit Pro is really for those customers who have got extreme needs,” said Tony Werner, Comcast’s executive vice president and chief technology officer.

The first 2Gbps Gigabit Pro deployments from Comcast will rely on Metro Ethernet that now serves Comcast’s midsized business customers, later migrating to passive optical network [PON] technology Comcast intends to begin installing in new housing developments and apartment complexes.

Multichannel News reports Comcast is now offering a limited promotional price of $159/mo for the ultra-fast broadband service, but to receive the discount customers must sign a three-year service contract with early cancellation penalties and agree to pay up to $1,000 in installation and activation fees.

Comcast claims it will offer the service to about 18 million homes by the end of the year — those within 1/3rd of a mile of Comcast’s existing fiber network.

Tony Werner, Comcast’s executive vice president and chief technology officer would not say which version of PON — GPON or EPON Comcast will use long-term, but a decision had already been made within the company and would eventually be known to customers.

ZTE, a Chinese provider of telecommunications equipment and network solutions, says EPON is the dominant fiber to the home solution in Japan, Korea, China and other Asia-Pacific countries. In other countries, especially in America, GPON is the preferred choice, as it can coexist with earlier PON systems.

Werner added Comcast will quietly deploy fiber to the home service in certain new housing developments.

“Once the trench is open, the incremental economics are close enough that we will do fiber-to-the-home, unless it’s a very small stub off of existing plant,” Werner said.

But for everyone else, it will be coaxial cable as usual unless customers pay that $1,000 fiber fee and are willing to wait up to eight weeks for installation. Comcast will run fiber and install the necessary equipment, including the Optical Network Terminal, only to customers who sign up for Gigabit Pro.

With video streaming now accounting for at least 64 percent of all Internet traffic, it should have come as no surprise to Australia’s ISPs that as data caps are eased and popular online video services like Netflix arrive, traffic spikes would occur on their networks as well.

It surprised them anyway.

Telecom analyst Paul Budde told the WAToday newspaper “video streaming requires our ISPs to have robust infrastructure, and to use it in more sophisticated ways, and that largely caught Australia off guard. I think it’s fair to say everybody underestimated the effect of Netflix.”

Not everybody.

Australia’s National Broadband Network (NBN) was originally envisioned by the then Labor government as a fiber-to-the-home network capable of enormous capacity and gigabit speed. Prime Minister Kevin Rudd proposed buying out the country’s existing copper phone wire infrastructure from telecom giant Telstra to scrap it. Instead of DSL and a limited number of cable broadband providers, the national fiber to the home network would provide service to the majority of Australians, with exceptionally rural residents served by wireless and/or satellite.

Conservative critics slammed the NBN as a fiscal “white elephant” that would duplicate or overrun private investment and saddle taxpayers with the construction costs. In the run up to the federal election of 2013, critics proposed to scale back the NBN as a provider of last resort that would only offer service where others did not. Others suggested a scaled-down network would be more fiscally responsible. After the votes were counted, a Coalition government was formed, run by the conservative Liberal and National parties. Within weeks, they downsized the NBN and replaced most of its governing board.

Netflix’s launch increased traffic passing through Australia’s ISPs by 50 percent, from 30 to 50Gbps in just one week, and growing.

Plans for a national fiber to the home network similar to Verizon FiOS were dropped, replaced with fiber to the neighborhood technology somewhat comparable to AT&T U-verse or Bell Fibe. Instead of gigabit fiber, Australians would rely on a motley mix of technologies including wireless broadband, DSL, VDSL, cable, and in areas where the work had begun under the earlier government, a limited amount of fiber.

In hindsight, the penny wise-pound foolish approach to broadband upgrades has begun to haunt the conservatives, who have already broken several commitments regarding the promised performance of the downsized network and are likely to break several more, forcing more costly upgrades that would have been unnecessary if the government remained focused on an all-fiber network.

Communications Minister Malcolm Turnbull has admitted the new NBN will not be able to deliver 25Mbps service to all Australians by 2016. Only 43 percent of the country will get that speed, partly because of technical compromises engineers have been forced to make to accommodate the legacy copper network that isn’t going anywhere.

Think Broadband called the fiber to the neighborhood NBN “a farce” that has led to lowest common denominator broadband. A need to co-exist with ADSL2+ technology already offered to Australians has constrained any speed benefits available from offering faster DSL variants like VDSL2. Customers qualified for VDSL2 broadband speeds will be limited to a maximum of 12Mbps to avoid interfering with existing ADSL2+ services already deployed to other customers. Only multi-dwelling units escape this limitation because those buildings typically host their own DSLAM, which provides service to each customer inside the building. In those cases, customers are limited to a maximum of 25Mbps, not exactly broadband nirvana. The NBN is predicting it will take at least a year to take the bandwidth limits off VDSL2.

The need for further upgrades as a result of traffic growth breaks another firm commitment from the conservative government.

NBN executive chairman Ziggy Switkowski told reporters in 2013 that technology used in the NBN would not need to be upgraded for at least five years after construction.

“The NBN would not need to upgraded sooner than five years of construction of the first access technology,” Switkowski said. “It is economically more efficient to upgrade over time rather than build a future-proof technology in a field where fast-changing technology is the norm.”

Since Switkowski made that statement two years ago, other providers around the world have gravitated towards fiber optics, believing its capacity and upgradability makes it the best future-proof technology available to handle the kind of traffic growth also now being seen in Australia. At the start of 2015, 315,000 Australians were signed up for online video services. Today, more than two million subscribe, with Netflix adding more than a million customers in less than four months after it launched down under.

Many ISPs offer larger data caps or remove them altogether for “preferred partner” streaming services like Netflix. With usage caps in place, some customers would have used up an entire month’s allowance after just one night watching Netflix.

But the online viewing has created problems for several ISPs, especially during peak usage times. iiNet reports up to 25% of all its network traffic now comes from Netflix. As a result iiNet is accelerating network upgrades.

Customers still reliant on the NBN’s partial copper network are also reporting slowdowns, especially in the evening. The NBN will have to upgrade its backbone connection as well as the last mile connection it maintains with customers who often share access through a DSLAM. The more customers use their connections for Netflix, the greater the likelihood of congestion slowdowns until capacity upgrades are completed.

Hackett

Optus worries its customers have extended Internet peak time usage by almost 90 minutes each night as they watch online streaming instead of free-to-air TV. Telstra adds it also faces a strain from “well over half” of the traffic on its network now consisting of video content.

This may explain why Internet entrepreneur and NBN co-board director Simon Hackett wishes the fiber to the neighborhood technology would disappear and be replaced by true fiber to the home service.

“It sucks,” Hackett told an audience at the Rewind/Fast Forward event in Sydney in March, referring to the fiber to the neighborhood technology. His mission is to try and make the government’s priority for cheaper broadband infrastructure “as least worse as possible.”

“Fiber-to the-[neighborhood] is the least-exciting part of the current policy, no arguments,” he added. “If I could wave a wand, it’s the bit I’d erase.”

Another cost of the Coalition government’s slimmed-down Internet expansion is already clear.

According to Netflix’s own ISP speed index, which ranks providers on the quality of streaming Netflix on their networks, Australia lags well behind the top speeds of dozens of other developed nations, including Mexico and Argentina.

But even those anemic speeds come at a high cost to ISPs, charged a connectivity virtual circuit charge (CVC) by NBN costing $12.91 per 1Mbps. The fee is designed to help recoup network construction and upgrade costs. But the fee was set before the online video wave reached Australia. iiNet boss David Buckingham worries he will have to charge customers a “Netflix tax” of $19.18 a month for moderate Netflix viewing to recoup enough money to pay the CVC fees. If a viewer wants to watch a 4K video stream, Buckingham predicts ISPs will have to place a surcharge of $44.26 a month on occasional 4K viewing, if customers can even sustain such a video on NBN’s often anemic broadband connections.

Some experts fear costs will continue to rise as the government eventually recognizes its budget-priced NBN is saddled with obsolete technology that will need expensive upgrades sooner than most think.

Instead of staying focused on fiber optics, technology the former Rudd government suggested would offer Australians gigabit speeds almost immediately and would have plenty of capacity for traffic, the conservative, constrained, “more affordable” NBN is leaving many customers with no better than 12Mbps with a future promise to deliver 50Mbps some day. There is little value for money from that.

Spanish consumers are switching to fiber to the home broadband service in droves as it becomes available around the country. In the last quarter, Spain tripled the number of fiber connections available the year before to a record 1,933,000 homes, according to a Spanish regulator.

Both DSL and mobile broadband options are losing interest with customers and have seen subscriber declines. Only cable broadband has grown alongside fiber, from 2,059,000 customers to 2,229,000 as of the end of March.

There are 12.83 million broadband lines in Spain, up from 12.14 million at the same time last year. Around 100,000 customers are signing up for fiber service each month. Most cite speed advantages fiber offers over competing broadband technologies.