As Craig Moffett readily admits, many on Wall Street view him as a cable analyst who jumped into telecom in 2007. In truth, however, Moffett is more of a telecom guy who took on cable in 2002 after joining Sanford C. Bernstein & Co. as a financial analyst.

For a 10-year period ending in the late '90s, Moffett was with Boston Consulting Group and devoted his time to consulting work at what is now Verizon — then Bell Atlantic and Nynex — as well as at various other global telecom giants, including BT, France Telecom and Telefonica.

So when Moffett says Verizon may never make a profit from FiOS and the telecom industry in general is falling behind cable in consumer broadband, he's not talking from personal bias but strictly looking at the numbers. And when he talks about the ascendancy of AT&T and Verizon in backbone networking, it's commentary worth hearing.

The broadband numbers — reflected in one of Moffett's latest reports, “U.S. Telecom, Cable & Satellite: A Subscriber Scorecard … Who's Winning the Wars?” — show that cable continues to gain overall subscription share and is gaining share in higher-margin voice and data businesses. Meanwhile, telcos are losing high-margin telephony customers in droves while attracting small, if growing, numbers of low-margin video customers.

Those video subscribers don't make up for the revenue lost to wireless substitution and cable voice over IP, however, because with each video subscriber comes the per-month cost for delivering content — paid to content owners — making the video business relatively low-margin, Moffett said.

“Not only do they incur all of the customer acquisition costs that are to be expected with any new venture … but you have to layer, on top of that, the cost of paying for content,” Moffett said. “And by all accounts, the telcos, by virtue of their limited scale in video, are paying very high monthly fees.”

By contrast, the cable companies' fixed costs already are covered in their previous network upgrades, which enabled them to deliver all three services — voice, data and video — over one network. “So the incremental margin for cable operators in adding voice or data is extremely high,” Moffett said.

For Verizon, the problem has been compounded by the heavy investment in its FiOS fiber-to-the-home network, which has consumed the company's capital, Moffett said.

“The problem facing Verizon is a simple one: You can't kiss all the girls,” he said. “They have to make choices about where they are going to deploy their capital, and they have made that choice emphatically in favor of FiOS.”

FiOS costs must be viewed in two ways, Moffett added: the shared network costs and the customer-specific costs. Verizon's shared costs are about $900 per home passed, he said, but if the company only signs up one-quarter of homes passed for video service, that raises the effective per-subscriber cost to almost $3600.

“Add to that the customer-specific costs of connections to the home, set-top boxes, customer-acquisition costs and promotions, and, in Verizon's case, a 19-inch HDTV,” Moffett said. “And on top of that, you have the ongoing cost of content.”

That brings per-customer costs to $4000, leading Moffett to conclude that FiOS is not a profitable proposition. “[Verizon's] returns on investment capital in wireline don't look interesting,” he said. “The investment they are making in FiOS gives them a world-class network that is, by all accounts, giving their customers a marvelous experience, but it doesn't look likely to ever earn an acceptable return on capital. The analogy is an auto company that is giving customers a Mercedes for the price of a Volkswagen. Customers are thrilled, but you don't make money that way.”

In addition, Verizon's investment is coming at a time “of significant uncertainty in the video world,” Moffett said, given Internet video.

AT&T's fiber-to-the-node approach to delivering IPTV also may not be profitable, Moffett said, but AT&T is spending less money and has done a better job of holding its own in data subscribers against the cable industry. That's why Moffett believes the Wall Street preference pendulum, which four years ago was certainly not in Verizon's corner but had swung in the pro-FiOS direction since 2006, is about to swing back to AT&T's less investment-intensive approach to broadband.

“Wall Street has begun to think that Verizon's future-proof network is the right strategy and that AT&T's U-verse is trying to put a Band-Aid on their underlying network problem,” Moffett said. “I suspect that the market may well reverse its view over the next couple of years, and AT&T's approach may regain favor.”

One reason, he added, is a very smart technology choice made by AT&T to bet on compression technology to squeeze more bits into its copper pipes and plan on copper-bonding and other technologies to continually raise the bandwidth bar.

“Unlike most of its competitors in the video business, who are locked into a specific compression algorithm by virtue of the chipset embedded in their set-top box, AT&T used software-based compression,” Moffett said. “That is a very important strategic advantage. It effectively means they have bet on the compression community, and that has been a pretty good bet in the past.”

And while Moffett's consumer broadband forecast looks gloomy for the telcos, he points out that this is one business segment for two mega-companies that have wireless operations, business services and their rich fiber backbone networks.

“This is a time of great promise for networks,” Moffett said. “You shouldn't read uncertainty of the video business as spelling demise of distribution as we know it — quite the opposite. Distribution is in a strong position at a time when content is facing these crosswinds of piracy on the Web and unwillingness of customers to view advertising and the tenuousness of monthly subscriber fees. Distribution is happily rising above all that because you have to pay for access to the cloud that is the Internet one way or another.”

In fact, Moffett sees another big pendulum swing — this one away from the edge of the network and back to the core.

“The center of the network had been so ruthlessly commoditized that a lot of people had given it up for dead, thinking that the value of the physical network had moved emphatically out to the edge,” he said. “Quietly but surely, the center of the network seems to be ascendant again. AT&T's and Verizon's OC-768 backbones once again will become sources of serious competitive advantage.”

That is happening because of the “triumph of the long tail,” Moffett said, as YouTube and similar sites make it much harder for content-delivery networks (CDNs) to cache popular content at the edge.

“If there are 10,000 out of a million Web sites that generate 95% of traffic, then CDNs are an ideal solution,” he said. “But if the top 10% of sites only generates 15% of the traffic, then you have a real problem on your hands, and there isn't an edge-caching mechanism that solves the problem. YouTube is that model in spades; it will make it awfully difficult for Google to make any money in the YouTube business because the cost of delivery is scaling with the demand. It may be that AT&T and Verizon make money off YouTube before Google does.”