Cisco’s Jeetu Patel: Telcos can’t ‘stay on the sidelines’ with AI

The industry is entering the “second big phase” of AI – a new, more consequential phase, according to Cisco’s chief product officer Jeetu Patel.

Moving beyond experimentation, he explained how service providers (SPs) will need to act to remain competitive. He argued that Cisco was in a great position to support SPs through this “pivotal moment for the industry” as the pace of innovation continues to speed up.

“You can’t stay on the sidelines … even when things get uncertain,” Patel said during an exclusive Cisco roundtable held at MWC Barcelona 2026. “Telcos need to seize the moment.”

Why AI could now change everything for SPs

The phase Patel describes is defined by a shift from AI as a productivity tool to AI as infrastructure – mainly with autonomous agents operating as “digital co-workers” that can handle complex tasks with minimal human input.

“This is an exciting opportunity,” he explained during the roundtable. “[Telcos] are going to need very different infrastructure. For every human worker, you’ll have 100-1,000 agents – that is close to one trillion agents working 24/7, which is a tremendous opportunity to provide services around connectivity, security and observability for the entire stack.”

Cisco, which Patel described as “the critical infrastructure company for the AI era,” provides low-latency, high-performance and energy-efficient networking. It provides optics and systems to make sure things can connect to data centres despite distance.

“We’re building our own silicon and chips, systems and operating system,” Patel added. “We are investing heavily in safety and security. There’s a lot of momentum and energy in the business, and we’re excited about what’s ahead.”

He added: “The sovereign AI movement is going to be real and sustained. Our ability to work with hyperscalers, private clouds, sovereign clouds, service providers and enterprises allows us to have one of the most comprehensive views in the industry.”

What rethinking infrastructure means for Cisco

The demand for AI workloads is already straining networks in ways that haven’t been anticipated. Given the technology’s infrastructure requirements, it is exposing gaps that legacy architecture wasn’t built to handle.

When it comes to Cisco, Patel said they are building technologies that help with operational efficiency in those areas. This has involved partnering with companies like Nvidia to drive greater efficiencies, in addition to Indosat and Jio.

Patel said these partnerships set them apart in the current market.

“This is an ecosystem play,” he explained. “No one company is going to be able to do it all; we have to work as a system.”

However, Patel was eager to highlight the importance of trust as the foundation of infrastructure development – supporting Cisco’s four pillars of time, trust, talent and technology.

“Everyone is very optimistic about the upside of AI, and everyone is very concerned about the downsides,” he said. “We have to make sure we approach this responsibly.”

When Capacity asked how Cisco would demonstrate its AI infrastructure is trustworthy, Patel said the company is building technologies that help with operational efficiency.

“The challenge most service providers have right now is: how do I keep this infrastructure up and running?” he said. “Only 20% of identified vulnerabilities ever get patched, and 80% go unpatched. Because upgrades are hard.”

He added: “We have to think about an entirely different way that infrastructure gets updated. That requires working with telcos, the public sector on policies and building technology that allows them to update infrastructure without taking it down.”

AI has reached ‘biggest transition’ point ever, tech leaders say

AI agents are starting to talk: Are the networks ready?