Cytranet’s CTO Doug Roberts has a favorite way of describing the last year in business connectivity: “Quiet revolutions,” he said. “When everything is working, it’s invisible. But when you upgrade the foundation—fiber routes, redundancy, monitoring, the way you deliver bandwidth—suddenly a business can do things it couldn’t do before.”
That “quiet revolution” is becoming louder for many organizations as AI tools, cloud workloads, and real-time applications strain networks that were designed for a different era. In an interview this week, Roberts described a shift he’s seeing across Cytranet’s customer base: more companies are treating internet performance as a strategic asset rather than a utility line item.
“Two years ago, a lot of conversations started with, ‘We just need faster internet,’” Roberts said. “Now it’s, ‘We need consistent performance, predictable latency, and an architecture that can handle AI-driven workflows, multi-cloud, backup replication, and video-heavy collaboration—without the network being the bottleneck.’”
### Why fiber is suddenly at the center of the AI conversation
It’s easy to think of AI as a software story, but Roberts argued that its most immediate impact is physical.
“AI adds more traffic, more east-west movement between systems, and more sensitivity to jitter and packet loss than people expect,” he said. “If you’re moving large datasets to and from cloud services, or syncing to offsite storage, or using real-time inference at the edge, your connectivity stops being a background concern.”
Roberts pointed to a growing pattern: companies aren’t only increasing speeds; they’re building resilience into the connection itself.
“Bandwidth is part of it, but reliability is the real headline,” he said. “We’re seeing increased interest in diverse paths, failover, and proactive monitoring so downtime doesn’t become a business event.”
### “Business internet” is becoming a design problem, not a purchase
Roberts described how the market has moved beyond a simple “megabits for dollars” comparison. In practice, what businesses need is a network design that matches their operational risk.
“There’s a difference between having an internet connection and having an internet strategy,” he said. “If a clinic can’t access patient systems, if a manufacturer can’t run an ERP, if a call center’s quality drops because of jitter—that’s not ‘internet being slow.’ That’s workflow disruption.”
He noted that many companies are now asking more sophisticated questions:
– How quickly can we fail over if a route is cut?
– Can we prioritize real-time traffic without complicated overhead?
– What’s our plan for scaling bandwidth when a new product launches—or when the team shifts to more cloud-first tooling?
Those are questions that lead naturally to fiber connectivity, he said, not only for speed but for long-term scalability.
“Fiber is the medium that gives you room to grow,” Roberts said. “It’s not just about today’s need. It’s about avoiding a network redesign every time your business evolves.”
### Datacenters are changing the definition of “local”
The other pressure point Roberts highlighted is where workloads live—and where they should live.
“Companies used to have this simple split: on-prem or cloud,” he said. “Now it’s much more blended. You might have cloud workloads, colocation in a datacenter, and on-site systems that need to talk to each other constantly. Connectivity is what makes that feel seamless.”
In that blended world, Roberts said, latency isn’t a niche concern. It affects everything from database performance to voice quality to the speed of nightly backups.
“Even when a workload is ‘in the cloud,’ there’s still a real network path from the user to that workload,” he said. “We spend a lot of time thinking about that path, and how to keep it stable.”
### The most newsworthy change: proactive networks
Roberts said the most meaningful change he sees in the industry is the move from reactive troubleshooting to proactive network management.
“Historically, internet service has been treated like power: you notice it when it’s out,” he said. “But businesses don’t want to wait for an outage. They want to know what’s happening before it becomes a problem.”
He described a push toward better telemetry, faster incident response, and more visibility into performance over time.
“What’s changed is the expectation,” he said. “Customers expect their provider to be watching the network like an operations team—because the internet connection is now a critical system, not a convenience.”
### What’s next: more capacity, more resilience, and fewer surprises
Asked what he expects to define the next phase of connectivity, Roberts didn’t focus on a single headline speed increase. Instead, he emphasized architectural improvements: redundancy, route diversity, and performance consistency.
“We’ll keep seeing higher throughput demands, but the bigger story is reducing surprises,” he said. “Businesses want fewer incidents, shorter incidents, and less uncertainty. When you’re running cloud apps, security tools, AI services, and collaboration platforms, the network has to be boring—in the best possible way.”
That “boring network” goal, Roberts added, is actually what unlocks the exciting part: organizations spending less time worrying about connectivity and more time building products, serving customers, and scaling.
“When the foundation is right, innovation becomes easier,” he said. “That’s what good connectivity is supposed to do—get out of the way and let the business move.”
