While I attended in participated in a panel and podcast at the Connected America Conference, one theme came through loud and clear: the network is no longer just infrastructure, it’s becoming an intelligent platform that enables, optimizes, and monetizes AI.
About a year ago, STL Partners published an article on How can telcos become more relevant enablers of AI. In the article, STL Partners contends that telcos need to begin adjusting their networks ahead of the growth in AI traffic. The premise is AI traffic will challenge conventional network designs, and if telcos take a ‘wait and see’ approach, they will deliver poor network experiences and miss the opportunity to capitalize on this growth.
As a Senior Business Development Manager at Cisco within the Americas Service Provider organization, I spend my time speaking with executives of service providers on modernizing their broadband infrastructure. After nearly three decades in the industry, I can confirm the STL Partners premise because the shift we’re seeing now, driven by AI, is one of the most significant yet. In my opinion, service providers that aren’t actively evolving their network to support next-generation technologies like AI, won’t just provide poor network experiences, their networks will be overwhelmed not only by traffic volumes, but also customer performance demands.
This isn’t an incremental change. It’s a fundamental re-architecture of how networks are designed, operated, and monetized. As STL Partners point out in their article, the need to grow edge facilities and prepare for East-West flows, is top of mind for my customers. As AI is adopted for their internal use and by their clients, the traffic patterns are new and putting strain in unfamiliar areas of their networks.
The Shift: AI Is Redefining Network Infrastructure
AI-enabled networks are forcing a move away from static, manually operated environments toward intelligent, adaptive systems.
1. Predictive, Demand-Aware Design
Networks are no longer built on assumptions; they’re built on insight. With Cisco’s automation, telemetry, and analytics platforms (such as Cisco Crosswork and Provider Connectivity Assurance and observability solutions), service providers can:
- Forecast traffic patterns and subscriber behavior
- Anticipate AI-driven workloads
- Optimize fiber routes and capacity in real time
This transforms network planning from reactive to predictive—reducing waste while improving performance.
2. Autonomous, Proactive Operations
Traditional networks wait for failure and require human intervention. AI-ready networks prevent failures by recognizing the early indicators and can take preventative actions ahead of time. Cisco integrates AI and machine learning into operations through platforms like:
- Cisco ThousandEyes (end-to-end visibility)
- Cisco Provider Connectivity Assurance
- Cisco AI-driven automation and assurance
These capabilities enable:
- Early anomaly detection before customer impact
- Root cause identification across complex environments
- Reduced truck rolls and operational costs
The result: networks that are self-healing, self-optimizing, and continuously learning.
3. Software-Driven Scale with Segment Routing
AI traffic is fundamentally different: bursty, unpredictable, and latency sensitive. And with the rise of agentic AI, it will only intensify. Cisco’s leadership in Segment Routing enables:
- Dynamic traffic engineering
- On-demand service creation
- Network slicing for application-specific performance
This allows providers to scale intelligently, not just massively.
Why This Matters: AI Traffic Changes Everything
We’ve already seen how streaming and gaming reshaped networks. AI will do the same, but faster and at greater scale. Unlike traditional traffic, AI workloads:
- Spike unpredictably
- Require ultra-low latency
- Demand consistent performance
That means networks must evolve from carrying traffic to understanding and adapting to traffic.
Cisco’s point of view is simple: AI shouldn’t just ride on the network; it should be built into it.
Monetizing AI: Turning Network Intelligence into Revenue
This transformation isn’t just technical, it’s economic. Service providers that embrace AI-ready infrastructure unlock entirely new revenue models. As STL Partners hinted at, telcos and providers that make this adoption in coordination with their customers, will earn the trust of those customers which can build a long-term profitable relationship.
Recently, I spoke with Lumen’s Distinguished Engineer Francis Ferguson on how they are adapting their network to support AI traffic and use cases. In this Beyond the Fiber “Voice of the Customer” podcast, we discussed the technical network changes required to support AI applications and customer expectations. The key points that Francis outlined were:
1. Outage Reduction = Immediate ROI
“In the customers’ experience, even if you get it right 99% of the time, that is still 100% wrong in the moment they need it and the network fails.” – Francis Ferguson
Using Cisco’s assurance and observability:
- Predict and prevent service degradation
- Reduce downtime and truck rolls
- Protect revenue while lowering OPEX
2. Smarter Capacity Planning
With AI-driven insights driven by telemetry, data models, and simplified topology, AI can build a digital twin that removes the guesswork of how to grow the network.
- Move from overbuilding to precision investment
- Optimize when and where to expand
- Improve ROI on every dollar of CapEx
3. Premium “AI-Ready” Connectivity
The market is shifting from speed to experience.
Providers can monetize:
- Guaranteed latency
- Low jitter
- Minimal packet loss
Target customers include:
- Gamers
- AI developers
- Enterprises running real-time applications
- Private networks that support critical networks for major industries such as Healthcare, Governments, and manufacturing
This creates higher-margin tier offerings beyond commoditized bandwidth.
4. Edge AI Services
Cisco enables distributed architectures that bring compute closer to the user. In the podcast, Francis highlights Lumen’s METRON (metro routed optical networking) and how this design simplifies their network and accelerates scalability for metros. (from 5:15 to 7:00 in the podcast)
Use cases include:
- Real-time video analytics
- Autonomous systems
- Smart cities and IoT
Benefits:
- New revenue streams
- Reduced backhaul costs
- Improved application performance
5. Network Slicing & Application-Aware Services
With Cisco’s programmable infrastructure:
- Allocate resources dynamically
- Create differentiated service tiers
For example:
- High-throughput lanes for AI training
- Low-latency lanes for inference
This allows providers to monetize performance, not just capacity.
Why Cisco: 40 Years of Building the Internet, Now Powering AI
The internet itself is over four decades old, and Cisco has been at the center of that evolution from the beginning. Today, Cisco brings that wealth of knowledge and experience to AI and offers a full-stack approach to AI-ready infrastructure:
- Access to Core Networking (Routed Optical Networking)
- Observability & Telemetry (ThousandEyes, Crosswork)
- Automation & Assurance (Provider Connectivity Assurance)
- Security & Data Sovereignty (Secure AI Factory)
- Edge and Cloud Integration (Compute, Unified Edge Platform, CX)
AI is no longer an add-on. It’s embedded across the portfolio, helping customers build networks that are not just faster, but smarter.
The Bottom Line: AI is moving from Connectivity to Intelligence
We’ve officially crossed a threshold; networks are no longer just connecting people; they are powering intelligence. The winners in this next era will be the providers who can:
- Measure experience
- Guarantee performance
- Monetize outcomes
The future isn’t about selling bandwidth. It’s about delivering and monetizing experience at scale.
To learn more about the Cisco technologies discussed above or in the podcast, please click here to go to our main solution page.
Additionally, the STL Partners study referenced above has been funded to include an in-depth discussion on how Telcos can support the AI revolution. It is being released as a three-part white paper that you can register to receive them here.
