- Technical debt at telcos has been boring – until now!
- Operators are concerned that their legacy infrastructure will hold them back from all the opportunities with AI
- Get the full story — download our free report for insights on AI in telecom networks
The 1996 presidential campaign for Bill Clinton used the mantra “It’s the economy, stupid” to keep the campaign focused. In the case of telecom, and not to be disrespectful to our readers, you could say "It's the legacy infrastructure, stupid." A big focus for operators right now is to clean up their technical debt so they can pounce on the opportunities in AI.
The focus on cleaning up technical debt and preparing legacy infrastructure for AI was evident in a Fierce Network Research survey conducted from April 18 to May 8, 2025, to gauge operators’ opinions about AI in telco networks.
When asked what the biggest challenge in adopting AI for network operations is, the largest group of survey respondents (32.11%) cited legacy systems and infrastructure, while 20.18% said data complexity and quality; and 19.27% said building a compelling business case, with smaller percentages citing other factors.
Recently, Fierce Network spoke with Joe Cumello, SVP and general manager of Ciena’s Blue Planet, who stressed the importance for service providers to clean up their inventory systems in preparation for agentic AI. Cumello said, “We have an expression at Blue Planet: You can’t AI what you can’t see. People are realizing when they apply an agent against bad data they get a bad outcome.”
In fact, Fierce Network has been calling attention to the problem of technical debt quite a bit, lately.
Scott Robohn, co-founder of the Network Automation Forum, has told Fierce, “If any telcos want to employ AI and train AI on their data, you need high integrity in the data you feed any model.”
One of the highest profile companies that has announced a concerted effort to clean up its technical debt is Lumen Technologies, which announced in January that it had hired the vendor OneVizion to clean up and organize its internal systems. Cumello said Blue Planet is also working with Lumen on its inventory cleanup effort.
What about AI for 6G networks?
Survey respondents for the recent Fierce Network Research report were also asked what are the top benefits AI will deliver for next-generation mobile and 6G networks, and 39.45% of telco executives answered that it was network optimization for enhanced performance, reliability and energy efficiency.
A couple of wireless executives who took the Fierce Network survey weighed in on this question.
Sylvestre Demonget, Verizon distinguished engineer for Network Strategy & Planning – Private Networks, said he answered network optimization for enhanced performance, reliability and energy efficiency. “In 6G, the RAN will embed AI for many/most of the RAN functions and features - sometimes referred to as ‘AI-RAN,’ which will bring self-optimizing networks to the next level; the best item to reflect this is probably the second bullet in my view,” said Demonget.
Vince D’Onofrio, VP of Networks for Rivada Networks, also shared thoughts about this survey question.
“This is a challenging question because there is no wrong answer, depending on one’s perspective,” said D’Onofrio. “The most compelling AI benefit, in my estimation, is spectral efficiency enhancement. This is because the network design and corresponding capital costs are largely driven by spectrum availability and efficiency. In large complex outdoor broadband networks, where most of the wireless network money is spent, interference issues are profound, ever changing and not well understood by humans without AI-enabled tools. It is the single most complex and costly issue that impacts all areas of the 5G cellular network.”
He said AI technology and information can be integrated into the network design process to preempt these problems by respecting the laws of physics and maximizing spectral efficiency.
Download the free report now: “AI and the Network: Telco leaders share insights on how AI is being used in the network.”
Other things you'll learn in our comprehensive report:
- Why AI network optimization is a stretch goal for telcos
- Why the telco edge is important for AI inferencing
- How AI might improve RAN performance
- Whether AI agents might soon be running the networks
- And more