BARCELONA – Lumen Technologies CEO Kate Johnson warned of the impact of digital sovereignty with regard to AI innovation, in an exclusive interview with SDxCentral held at this year’s edition of Mobile World Congress (MWC).

In a sit-down chat, the Lumen leader gave an overview of the vendor’s network roadmap, as well as a message for the telecom crowds at MWC.

“The AI economy needs a programmable network,” Johnson said. “The network has become the supply chain for data, and it needs to be agile and secure and high bandwidth, and it needs to be connected to everything so you can move data from anywhere to anywhere.”

Touting Lumen as already having built this architecture, Johnson added the term telco is a misnomer in this scenario.

“I think the companies that want to be digital network services companies understand the importance of an API-driven, programmable platform that interconnects with everything, and that this is an ecosystem play. It's a page out of the technology world's book," Johnson said. "So I think that there's an awakening of this happening, and it represents an enormous opportunity for companies that haven't innovated in a while: those legacy telco companies.”

Obviously, Johnson knows a lot about the tech world playbook. The exec’s career history includes president roles at Microsoft, Oracle, and Red Hat. When asked how that history comes into play at Lumen, Johnson said everything she learned in her time at Microsoft is being applied to the company.

“I often say, ‘hey guys, we've seen this movie before,' when compute and storage were cloudified right back in, I guess ... the first offering was officially 2006 from Amazon," Johnson said. "But when Microsoft got in the game in the early teens and then I joined in 2017, it was gas to the pedal time where we got to exponential growth and had to change everything.”

The CEO remembers changing all and sundry, from the offerings and pricing portfolio at Microsoft, to the consumption model and the way it was sold, supported, and marketed.

“Companies had to change the way they addressed risk," Johnson said. "They had to shift from capex to opex. The way that they purchased it was different, and the way that they paid for it was different, and they had to think about consumption for the first time.”

As such, Johnson is seeing history repeat in the network space, as the network is cloudified with reverberations felt in the same business areas.

“So everything that I learned, we're just applying and actually accelerating it," Johnson said. "And the connected ecosystem, that is how technology companies make lots and lots of money, right? And that is why I brought that to Lumen, because it is an opportunity to extend our commercial rate reach and hit the gas pedal on growth.”

Borders and bubbles

That growth correlates with the AI opportunity: Lumen has previously expected the next wave of AI market pressure to come from AI model inference, with demand led by financial services, retail, and health care.

Dell'Oro research found that global data center capex increased 51% in 2024, hitting $455 billion in investment thanks to custom accelerators from the cloud giants; this was fueled by the initial stages of AI training, the precursor to the much more capacity-heavy strains of inference workloads. In other words, the market is moving from baby steps, which were expensive enough; IDC recently predicted edge computing spend alone will grow to nearly $378 billion by 2028, thanks to inference, while Amazon Web Services (AWS) claimed to this title that inference is already a multimillion-dollar business for its cloud wares.

What may halt this expansion beyond the supply chain specter is the sovereignty question. Johnson stressed that while she’s a “huge fan” of data sovereignty, she has some concerns on digital sovereignty.

“What worries me about digital sovereignty is this notion of keeping it closed," Johnson said. "I understand the need for data sovereignty and data privacy when it comes to thinking about the entire stack being walled off. But I get a little concerned about the pace of innovation, and it goes against what everybody's trying to do is build it as fast as they can and then be a net exporter of AI services.”

Lumen’s chief recalled her tenure at Red Hat, where she learned the power of open source setting the pace of innovation as “materially faster” than if it were closed or proprietary.

“I think if we can find a way to instead govern it, cross borders and actually share innovation, I think it's going to be a net positive for humans and for business," Johnson said.

Johnson is less concerned about the AI bubble, echoing similar sentiment at MWC from the likes of Cisco that, instead of overestimated demand, there is a serious danger of underestimating what’s needed to build out neo-networks for the AI era.

“From my perspective, networking is critical infrastructure. And ... infrastructure is boring until it's critical, and then it's an emergency," Johnson said. "And so what we're seeing right now is a bandwidth reckoning where these big companies are saying, ‘oh my gosh, I got to upgrade.’ Then it works like, how fast can you go? And we're using our existing conduit and pulling through these state-of-the-art fiber solutions to sometimes quadruple existing rounds of bandwidth, and we just can't seem to go fast enough to meet the need, which is very exciting for this company.”

Lumen is thus continuing with its biggest ever expansion, seeing an already enormous amount of traffic courtesy of AI agents in spite of how enterprises are still in the nascent stages of deploying AI at scale.

“When I started back in November of 2022, we had 12 million fiber miles for our intercity network, and by the end of 2031, we will have 58 million fiber miles," Johnson said. "It's the largest expansion of the internet in history, and it's all powered by the hyperscalers and the neoclouds and the social platforms doing the private connectivity fabric deals with us for this massive upgrade.”

The light of Lumen 'delights'

Johnson added that with questions around AI’s impact on employment and bona fide productivity gains, companies are already taking their time instead of recklessly making deals left, right, and center.

“You can predict the likely outcomes before you write very large, multibillion-dollar checks," Johnson said. "I think that the companies that are writing those very large checks are looking at things from a vantage point of deep knowledge, and so the investments that they're making in infrastructure to support the development of AI services, I trust that they understand what's required to do that.”

The twist, in Johnson’s eyes, is that some businesses are being myopic when looking at AI from a productivity viewpoint.

“I think companies that focus on productivity are going to have temporal or short-term gains. And I think companies that think about using AI to delight customers, I think their value is going to be enduring," Johnson said. "That's really what business is all about: customer obsession."

When asked how Lumen is using AI to delight clients, Johnson pointed to improved customer service and sales opportunities, and its potential for autonomous, self-healing networks.

"We use it in a bunch of different ways. And frankly, we're imagining the art of the possible, with what an identified digital network services company looks like. And the way that we do that is we put the customer at the center of everything," Johnson said. "How are they going to want to purchase, deploy, provision, and support their networks in the world of AI? Predictive maintenance has been around for a while, but now it's getting pretty sophisticated because you can have an outage and have it fixed before anybody ever knew there was an outage.

"So that's what we're thinking about, how we can use AI to fundamentally delight our customers. Because then they'll keep coming back to us, and that's the name of the game," Johnson added.