- The broadband industry has spent the better part of a decade talking about speed
- The next major transformation may have surprisingly little to do with bandwidth at all
- The real breakthrough will be intelligence
The broadband industry has spent the better part of a decade talking about speed. DOCSIS 4.0. Multi-gigabit services. Symmetrical bandwidth. Faster upstreams. Higher capacity.
Those upgrades matter. But after speaking recently with Todd McCrum, SVP and GM of the Broadband Access Business Unit at Applied Optoelectronics (AOI), I came away thinking that the industry's next major transformation may have surprisingly little to do with bandwidth at all.
The real breakthrough will be intelligence.
For decades, cable operators have operated vast outside-plant networks with only limited visibility into what was happening in the field. They could see symptoms. They could infer likely causes. But when something went wrong, the process often resembled detective work. A customer would call. A truck would roll. Then came the hunt: utility poles, railroad rights-of-way, bridges, subway tunnels, river crossings and household wall cavities, all potential hiding places for the fault.
As McCrum described it, operators frequently knew there was an issue but still had to hunt through multiple network segments and dozens of potential failure points to find it.
"It's like a needle in a haystack," he told me.
What AOI has developed is not simply another bandwidth upgrade. It is a way of bringing telemetry and operational intelligence from active devices throughout the network back to a central location where that information can be analyzed and acted upon.
That may sound incremental.
It isn't.
In fact, McCrum believes it addresses a challenge the cable industry has been trying to solve for decades.
"I've seen the industry trying to do this for more than 40 years," he said.
Historically, operators had relatively good visibility into centralized infrastructure such as headends and hubs. The further networks extended into the field, the more opaque they became. AOI's approach changes that equation by collecting health and operational data from active devices throughout the network and making it available in real time.
The immediate benefit is obvious. Instead of dispatching multiple crews to search for a problem, operators can identify precisely where the issue exists and send the right person to the right location.
"Let's not send two or three or four bucket trucks out there and try to figure it out," McCrum said. "We can tell you exactly where the issue is and send someone there immediately."
Building the knowledge base
That alone creates significant operational value.
But the bigger opportunity emerges over time.
As networks accumulate years of telemetry data, operators begin building something much more valuable than a monitoring platform. They begin building institutional memory.
This is where McCrum's view of artificial intelligence becomes particularly interesting.
Unlike many technology executives, he is not talking about large language models running the network. He is not describing autonomous agents making unsupervised decisions. Instead, he talks about machine learning in its most practical form: identifying patterns, correlating events, and recognizing conditions that historically precede failures.
If a device consistently operates above its normal temperature range, for example, the system can recognize that pattern and flag it long before customers experience an outage.
"I've been monitoring this for 48 hours," McCrum said, describing the type of insight the system can generate. "My history of logs tells me when this happens, something's not seated right or we have a power problem."
That may sound mundane compared with some of the grand claims currently being made about AI. In practice, it could prove far more valuable. It is, bluntly, the type of pragmatism that our industry is overdue for.
The broadband industry does not need hallucinating software agents making decisions about network operations. It needs systems capable of processing millions of data points, identifying anomalies, and presenting meaningful recommendations to human experts.
Importantly, the human remains in control.
That point surfaced repeatedly during our conversation, particularly when the discussion turned to jobs. One of the most persistent fears surrounding AI is that it exists primarily to replace people. McCrum sees it differently.
"It's letting people do the job they were actually hired to do," he said.
Instead of spending their time chasing faults after customers have already been affected, technicians can focus on preventing problems before they occur. The work shifts from reactive maintenance to proactive maintenance. The objective becomes improving network health rather than simply repairing network failures.
That distinction matters because it points toward a broader transformation that extends well beyond operational efficiency.
The real value may not be fewer truck rolls. It may be better customer experiences.
For years, operators have competed primarily on speed. Increasingly, however, reliability is becoming just as important. Consumers expect broadband to be available all the time. Work, education, entertainment, communications, and increasingly AI services all assume continuous connectivity.
As McCrum observed, when networks fail today, customers do not simply shrug and wait for service to return.
The expectation is constant availability.
That raises the stakes considerably.
Faster repair times help. Preventing failures altogether helps even more.
Perhaps the most intriguing aspect of AOI's vision is the possibility that the system becomes smarter as more operators contribute data. According to McCrum, providers are generally willing to share information about root causes and failure patterns, provided individual operator data remains confidential.
The result is a growing operational knowledge base capable of recognizing recurring conditions across the industry.
"We've seen 80,000 instances that manifest in this way," he explained. "In 70,000 of them, this was the root cause."
That is a remarkably powerful concept.
Every repair becomes a learning event. Every failure contributes to a larger body of operational knowledge. Every technician effectively trains the system for the next technician.
Viewed through that lens, AOI is not simply building monitoring tools. It is helping create a collective intelligence layer for broadband infrastructure.
The timing could prove significant.
As operators prepare for DOCSIS 4.0 deployments, network quality becomes increasingly important. Higher speeds demand cleaner plant conditions and tighter operational tolerances. Telemetry allows operators to identify weak points before upgrades occur rather than discovering them after launch.
"You can look across your whole network and run these algorithms," McCrum said. "Where am I going to have problems deploying this?"
That capability alone could save operators enormous amounts of time and money.
But I suspect the larger story lies elsewhere.
The more I listened, the less this felt like a DOCSIS story.
It wasn't really an AI story either.
It was a data story.
For decades, broadband networks have primarily been transport systems. What AOI is describing is something fundamentally different: a network that continuously observes itself, learns from itself, and improves itself.
In effect, the outside plant begins to develop something approaching operational awareness.
The network stops being something operators simply run.
It becomes something they understand.
And that may ultimately prove to be a far more important upgrade than another few gigabits per second.
Stephen M. Saunders MBE is a communications analyst and USPTO-registered inventor examining how digital infrastructure — 5G, cloud, and AI — is reshaping industry, power and society, as well as underpinning the emerging, ubiquitous global digital economy. As anchor of FNTV and a longtime industry insider, he focuses less on growth narratives and more on execution, risk and how hyperscale technology is distorting markets, governance and society at scale.
Opinion pieces from industry experts, analysts or our editorial staff do not represent the opinions of Fierce Network.