Cytranet’s CTO Doug Roberts sees a broadband “reset” coming — and it’s built on fiber, smarter networks, and AI-ready data routes
For years, the conversation about connectivity has been dominated by speed: gigabits, low latency, bigger pipes. But Doug Roberts, CTO of Cytranet, says the next chapter is less about raw numbers and more about what businesses can reliably do with their connections — especially as AI tools move from pilots into everyday operations.
“People don’t just want faster internet,” Roberts said in an interview. “They want confidence. Confidence that their network will perform the same way at 10 a.m. on a Monday as it does at 10 p.m. on a Saturday. And they want it to support new workloads — AI inference at the edge, cloud connectivity, real-time collaboration — without needing a total redesign every year.”
That shift is arriving as fiber connectivity expands into more commercial corridors and secondary markets, while data center footprints grow to support everything from cloud platforms to specialized AI infrastructure. Cytranet, Roberts explained, is positioning its business internet strategy around this convergence: fiber access paired with the operational rigor businesses expect when connectivity becomes as critical as power.
A new kind of “AI readiness” for business internet
Roberts described a noticeable change in what customers are asking for. Even companies that don’t think of themselves as “AI companies” are deploying AI-enabled workflows: voice transcription, intelligent document processing, customer support automation, analytics, and security monitoring.
“AI changes traffic patterns,” he said. “It’s not always a steady stream like classic web browsing. It can be bursty, it can involve large uploads, and it can require fast, consistent response times. When that starts happening across an organization, a connection that ‘usually works’ isn’t good enough.”
According to Roberts, the most important questions have become practical ones: how quickly an issue is detected, how predictable performance is during peak hours, and how well traffic can be routed to cloud on-ramps or data centers when applications are distributed.
He pointed to latency consistency as an underappreciated factor. “A lot of people focus on average latency, but what breaks real-time apps is variation — jitter — and congestion events. If you’re running voice, video, point-of-sale, remote desktops, and now AI-driven tools, you need a network that behaves consistently.”
Fiber buildouts meet the data center boom
The past two years have seen growing investment in data center capacity and interconnection as cloud demand rises and AI accelerates the need for compute. Roberts said that, for business customers, the data center conversation isn’t abstract — it directly affects where their applications live and how quickly users can reach them.
“Businesses are more distributed than ever,” he said. “Their people are spread out, their apps are in multiple clouds, and their customers are everywhere. So the question becomes: how do you get from your office to the workloads — securely, quickly, and with minimal points of failure?”
Roberts emphasized that fiber connectivity matters not just because it’s fast, but because it’s stable and scalable. “When you’re on a fiber-fed network, you can plan for growth. You can add bandwidth without swapping out the entire access method. That’s huge for companies that don’t want disruption every time their needs change.”
But he cautioned that fiber alone doesn’t solve everything. “You can have a fiber circuit and still have a poor experience if the upstream routing is messy or if there isn’t enough redundancy. The real story is how the network is engineered end-to-end.”
What “reliability” means in 2026
Asked what reliability looks like now, Roberts talked less about marketing terms and more about operational habits: proactive monitoring, clear escalation paths, and designing for failure instead of assuming it won’t happen.
“Networks fail. Backhoes happen. Hardware fails. Software has bugs,” he said. “The difference is whether you’ve designed for that reality. Do you have diverse paths? Do you have monitoring that tells you there’s a problem before the customer calls? Do you have people who can actually fix it quickly?”
Roberts said more customers are asking about redundancy not as a luxury, but as a baseline. Some are pairing primary fiber with a secondary fiber route or fixed wireless backup. Others are asking for diverse entry points into buildings.
“It used to be that redundancy was only for large enterprises,” he said. “Now, a mid-sized company running cloud-based phone systems and remote work can’t afford an outage. Their entire operation is the internet.”
The small details businesses are paying attention to
One of the more striking parts of the conversation was how specific customer questions have become — a sign that IT teams are under pressure to justify spending while also preventing downtime.
Roberts said customers increasingly ask:
• How quickly can a circuit be installed, and what causes delays?
• What’s the process for troubleshooting if performance degrades?
• What does “symmetrical bandwidth” mean for their actual applications?
• Can the network support secure site-to-site connectivity for multiple locations?
He noted that businesses are also more skeptical about vague promises. “They want transparency. If there’s a maintenance window, they want to know. If there’s an outage, they want timely updates. Trust comes from communication, not just a contract.”
How AI is influencing network operations
Roberts also discussed AI from the provider side. While he avoided hype, he acknowledged that AI-assisted tools are improving how networks are monitored and maintained.
“There’s value in using machine learning to identify anomalies in traffic patterns,” he said. “If you can detect early signs of congestion or failing optics before it becomes a full outage, that’s a better customer experience. But it has to be paired with good engineering. AI can’t compensate for poor design.”
He framed AI as a way to reduce mean time to repair and to prioritize issues more intelligently, especially when providers manage many links across diverse infrastructure.
The bigger takeaway: connectivity is becoming infrastructure again
In the last decade, internet access often felt like a commodity purchase: pick a speed tier and move on. Roberts believes that mindset is fading as broadband becomes the foundation for business continuity and growth.
“Connectivity is infrastructure again,” he said. “It’s like electricity. You don’t think about it until it’s gone — and by then it’s too late. Businesses are realizing they need the same seriousness in how they buy and manage internet service as they do with any critical system.”
Roberts thinks the most successful organizations over the next few years will be the ones that treat connectivity as a strategic asset rather than an afterthought — particularly as AI tools and data-heavy workflows become normal.
“If your network can’t keep up, your business can’t keep up,” he said. “The good news is, with fiber expansion and better network practices, we can build connectivity that’s not just faster, but genuinely dependable.”
