At Mobile World Congress, Cisco showcased paths to modernize for AI, monetize new revenue streams and move from the backbone to the brain of the digital economy.
In a world where network capacity no longer equals cash, communication service providers (CSPs) are racing to reinvent themselves for the AI era — and the clock is ticking.
For decades, carriers have been the largely invisible backbone of the internet, focused on scale and reliability. That posture no longer cuts it. AI is already reshaping traffic patterns and economics. Global data will triple by 2028. And over the next 10 years, two-thirds of all internet traffic will be AI-related, Cisco experts say.
That kind of growth puts real pressure on current architectures — and on CSPs’ ability to monetize what runs across their pipes. At the same time, AI workloads are moving from centralized data centers to the edge, where decisions need to happen in real time for use cases like robotics, industrial automation and smart cities.
“This is the next big evolution,” says industry analyst Jeff Kagan, who contends that the service provider industry faces a unique challenge and while some will meet the challenge, “every one of them is going to struggle with it,” he continues. No one knows how quickly everything is going to change, only that it will change, he adds.
Legacy networks once built for stability now must adapt for agility.
At Mobile World Congress (MWC) 2026 in Barcelona in March, Lumen CEO Kate Johnson highlighted how AI requires rethinking network infrastructure, noting that enterprise data demands are forcing the redesign of infrastructure to be faster and more programmable.
“The architectures are changing, the workloads are exploding, and users are getting more impatient,” said Johnson, who added that many companies are discovering their existing networks are “not big enough, fast enough, secure enough or smart enough” to support AI-driven computing.
Jeetu Patel, president and chief product officer at Cisco, notes: “The pace of innovation in AI is astounding. Technical breakthroughs are just beginning to translate into new experiences for consumers and applications for businesses that will reshape how the world works and connects.”
Patel emphasizes that this shift “presents massive opportunities for service providers to grow their businesses and reduce costs by modernizing their infrastructure for AI.”
Speaking with theCUBE’s John Furrier at MWC, Patel said of telcos’ excitement for AI: “As the industry moves to this second phase of AI from chatbots to agentic, the entire infrastructure is going to have to be rethought because these agents are going to be working 7x24, they’re going to be very data hungry, they’re actually going to be talking to each other.”
He added that the ability to interact and connect with the data center where digital workers live and the campus branch where human workers are “will be done through a secure global connectivity fabric, which is essentially the service providers.”
“They will have a whole different opportunity to monetize. They’ll be able to have tiered services. They’ll be able to make sure they can deliver a degree of sophistication in the service that’s going to be needed for the next wave.”
Service providers seeking to monetize these new revenue streams are looking to Cisco’s Agile Services Networking, which delivers the silicon, systems and software innovation they need to thrive.
The end-to-end architecture offers a unified architecture for 5G, connecting access, edge and core— without ripping out existing infrastructure.
Edge: Where AI gets real
CSPs increasingly recognize that the edge is where AI promises real business outcomes — and where the integration pain shows up first.
“The edge is really the network. It’s where operational decisions are made,” explains Devin Yaung, SVP, Group Enterprise – IoT Products and Services at NTT. “Things that have low latency, that are really need immediate action have to be done at the edge. Is this machine about to fail? Is someone in danger in a worker safety zone? How much inventory do I currently have? These things need to be processed at the edge.”
But stitching together that edge fabric is far from simple.
“The question for the industry is what’s that edge infrastructure service that we can provide that solves problems and fosters innovation, and what that’s ultimately going to be is a shift in how we think about mobile computing,” says Leonard Lee, founder and executive analyst at neXt Curve.
The challenge is forcing providers to be thoughtful about how they approach modernization and how they support the future of AI, Lee adds.
Cisco’s answer is Agile Services Networking — integrating networking, security and compute into validated, repeatable designs that operators can use to stand up edge services faster. The idea is to mask complexity behind a platform that works across fiber, private 5G, Wi‑Fi and partner ecosystems.
For CSPs trying to turn “upload bandwidth” and underused spectrum into revenue, that kind of packaging could matter.
No‑rip modernization play
The other big message from Cisco: telcos don’t have to start over. With Cisco Agile Services Networking, operators can overlay AI‑ready capabilities on top of existing networks, then expand as demand grows.
A key pillar is Cisco Silicon One, the company’s unified, high‑speed silicon platform designed for AI‑heavy networks.
Nvidia selected Silicon One as the only other silicon to be standardized as part of Nvidia’s Spectrum-X Ethernet networking platform, making it the only third-party silicon standardized for this specific AI-accelerated networking platform.
Silicon One, paired with Cisco optics and systems, is aimed squarely at the surge in east‑west traffic between GPUs and switches. For operators, that’s about driving density, power efficiency and performance in the parts of the network where AI workloads live.
Partnership as the differentiator
Technology alone won’t get telcos over the line. Several of Cisco’s carrier partners emphasized that co‑innovation and integration are equally critical.
Bart Janssens, principal network architect at Colt Technology Services, a user of Cisco Agile Services Networking, says customers expect faster service delivery. Cisco’s partnership model is what allows Colt to keep evolving its network for AI use cases.
“Everybody has a moment of growth. What is interesting with Cisco is that the moment comes again and again. What I like about Cisco is that it's a partnership. It’s not just about big building blocks; it’s about listening to the customer and working on specific needs. It’s not so much about the boxes; we really like the glue to do more with those boxes.”
That “glue” — validated designs, integration support and joint go‑to‑market — is the kind of help many CSPs will need as they move from connectivity providers to AI platform players.
From backbone to brain
Taken together, Cisco’s message out of MWC is clear: AI is rewriting the economics of networking, and service providers have a narrow window to decide whether they will simply carry AI traffic — or monetize it.
Those that lean into AI‑ready architectures, edge‑focused solutions, and deep vendor partnerships could move from the backbone of the internet to the brain of the AI economy. Those that wait risk watching value flow — once again — to players sitting on top of their networks.