Lumen CMO on cloudifying telecom: fiber demand tripling as AI agents generate unseen data volumes

Summary

Lumen is positioning connectivity as the missing infrastructure layer in the AI economy, arguing that compute, storage, and cooling have absorbed most of the capital conversation while bandwidth demand quietly maxes out existing capacity. Ryan Azdorean, Lumen's Chief Marketing and Strategy Officer, making the case at Palantir's AIPCon on September 4, 2025, put the company's fiber build-out in concrete terms: by 2028, Lumen will operate roughly 66 million route miles of fiber, representing three to five times prior capacity levels.

The AI Bandwidth Argument

The intuitive objection — that AI workloads are just text packets, nothing like bandwidth-intensive video streaming — misses where the data actually originates. The volume is not driven by user prompts but by the machine-to-machine traffic those prompts trigger. A single query to a model running deep research fans out to dozens of external calls; enterprise systems running real-time scheduling, inferencing, and planning at companies like American Airlines and BP generate continuous background traffic that never appears in a chat window.

Azdorean's framing is that AI agents are the demand driver that legacy telecom infrastructure was never sized for, and that this shift is structural rather than cyclical.

Cloudifying the Network

Lumen's commercial strategy centers on what Azdorean calls "cloudifying telecom" — replacing the historical one-port-to-one-service model, which required a physical truck roll for every new circuit, with a software-defined, self-service abstraction layer. Customers get a portal through which they can provision bandwidth, select locations, and adjust speeds on demand. The pricing model shifts accordingly, from fixed subscriptions to consumption-based billing that accommodates the spike-and-trough patterns common in sports, manufacturing, and healthcare.

The multi-cloud angle is increasingly central. Enterprises that once built on a single cloud provider are now hitting Azure, GCP, and AWS simultaneously across multiple US regions, and they want a unified fabric rather than discrete dedicated connections to each. Lumen is pitching itself as that interconnect layer.

Palantir's Role

The Palantir partnership targets Lumen's internal operational complexity, specifically the accumulated software debt and legacy systems layered over decades of network buildout. Palantir is being used to rationalize those systems, automate network optimization, and build toward a model where enterprise customers receive AI-managed network performance without configuring it themselves. Azdorean frames this as the most commercially significant part of the partnership, not just a modernization exercise but a prerequisite for selling network-as-a-service at scale.

Lumen enters this buildout cycle carrying significant debt from prior years, and the credibility of the demand narrative — that AI agent traffic will sustain a three-to-five-times capacity expansion — remains a key variable investors will need to pressure-test independently.