Charter CEO says the idea that fiber is superior is ‘just dead wrong’

At this week’s MoffettNathanson Media & Communications Summit in New York, Craig Moffett asked Charter’s CEO Tom Rutledge a question that’s been on a lot of people’s minds. He asked if cable operators will inevitably have to spend big bucks on fiber deployments to stay competitive in the broadband business.

Specifically, Moffett asked, “How confident are you that the HFC plant really can stay competitive with fiber?”

Rutledge said, “We have a lot of fiber in our network, and it’s really a question of where do you end the fiber, and what technology do you use to maximize the connectivity with the end device?”

He noted that even a fiber feed directly to a house doesn’t deliver fiber to a device. What delivers connectivity to a device is actually Wi-Fi in most cases.

Charter has built an advanced Wi-Fi capacity, which is backed by a fiber backbone and uses hybrid fiber coaxial (HFC) cable in the local neighborhood at short distances. A fiber-optic cable doesn’t connect directly to a device in the home.

“The idea that this technology [fiber] is transformative and superior is just dead wrong,” said Rutledge. “It’s just another form of transmission.”

There would be many in the fiber broadband camp who might disagree with that assessment. They’re lobbying the NTIA heavily to spend the billions of dollars from the Infrastructure Investment and Jobs Act (IIJA) on all-fiber deployments.

In any event, Rutledge laid out the template for Charter’s plan to evolve its wired network.

Charter is taking a high-split path to DOCSIS 4.0.

Dell’Oro Group VP Jeff Heynen explained that “high-split” means that a cable operator is reallocating the use of its spectrum. “There’s only so much spectrum,” said Heynen. “Historically, more was allocated to the downstream. But with Zoom and gaming, you’ve got to allocate more upstream. So high-split is basically allocating more spectrum to the upstream side.”

High-split is a DOCSIS 3.1 technology. Heynen said operators might have to swap out some amplifiers in their cable plant, but the whole thing is an effort that’s not super expensive for them to do.

Rutledge also said, “In the high-split technologies that we’re deploying, we don’t think we have to go to Node plus 0.”

Heynen explained that Node plus 2 or Node plus 3, etc, indicates the amplifier cascades that sit beyond the node. “A typical node is plus 4 or plus 5,” said Heynen. “The further you pull fiber that’s how you get to Node plus 0. But the cost to pull fiber to remove all those amps is very expensive.”

He added that “in high-split it is possible to move from Node plus 5 to Node plus 2 or 3.”

Rutledge said, “The active electronics in an all-fiber network aren’t as low as you think because you have a lot more nodes. I was shocked when I first did the analysis of how little electronics you actually save. Even with DOCSIS 4.0, which is much higher capacity, we don’t think we have to go to Node plus 0.”

Heynen said that Charter might also need new customer premise equipment (CPE). Early DOCSIS 3.1 didn’t support the upstream frequency ranges for high-split. “You still have probably a legacy base of DOCSIS 3.0 and 3.1 CPE that you’ll have to swap out,” he said.

The cost of upgrading

In terms of the cost of Charter’s upgrade path to DOCSIS 4.0, Rutledge said, “The upgrade path that we have from a capital investment perspective is very efficient. We went to 1 Gig service by doing the DOCSIS 3.1 rollout, which costs $9 a passing. The next upgrade – the spectrum split upgrades we’ll be doing going forward will cost more than that. But not nearly the kind of costs that it takes to build a new network.”

Of those companies that are embarking on all-fiber deployments, Rutledge said, “I think it’s very difficult to get a return. And the higher the cost of capital the quicker you find that out.”