As AI reshapes infrastructure, Cisco responds to a changing data center
Cisco was founded 41 years ago as a network equipment company. While it has dipped into a number of other technology areas over the years through an aggressive acquisition strategy, its core networking mission has remained steady — even as networking itself has evolved from TCP/IP to Ethernet to WiFi to cloud-era, software-defined networks. None of those shifts, however, has been as significant as the impact of generative AI on networking.
The scale and speed required to run AI workloads are changing how data centers are built and operated. That shift puts a substantial amount of money on the table for a company like Cisco, and chief product officer Jeetu Patel clearly recognizes the opportunity.
"Cisco has fundamentally changed quite dramatically over the course of the past 15 to 18 months. I would say we've probably done more innovation in the past 18 months than we've done in the previous decade combined," Patel told FastForward. (Note that I spoke to Patel in November.)
Patel sees this as a pivotal moment for Cisco, and for the tech industry more broadly. In practical product terms, that strategy has translated into a wave of new routers, switches, and GPU-ready servers optimized for AI workloads, along with new AI chips released just this week. It has also reshaped Cisco’s data and security strategy, most visibly through the Splunk acquisition and a renewed focus on AI-driven cybersecurity.
Where they're headed
Prior to the AI boom, the networking sector was slowing considerably as post-pandemic demand subsided. Holger Mueller, an analyst at Constellation Research, says that the changes we're seeing from AI create an opening for everyone.
"AI gives all data center suppliers another shot to stay relevant. For Cisco the goal has to be to provide its networking gear into cloud data centers as the on‑premises business shrinks," he said. "The other opportunity is AI‑powered networking software, with IT operations (ITOps) and cybersecurity options."
Patel says that the company is seeing interest from a wide range of customer types. "We're engaging with companies with major build-outs all over the world," he said. That includes hyperscalers, neoclouds, sovereign clouds, service providers and enterprises.
Structural changes
To take full advantage of these new market opportunities, the company required an internal structural change and a commitment from the very top, CEO Chuck Robbins, to make substantive changes to the way the company operates.
About two years ago, around the time that Cisco was finalizing the $28 billion Splunk deal, Robbins decided to consolidate the entire product line under Patel's leadership. "Chuck came to me and said, 'I'm not just going to give you Splunk, I'm going to give you the entire organization, and we've consolidated all of product under one leader for the first time in a very long time at Cisco.'"
~Holger Mueller, Constellation Research
It wasn't just a vote of confidence in Patel's leadership. It also marked a way to rethink the entire company with faster decision making, something that is essential in the quickly shifting AI landscape. Companies have to move faster, and they can't do that with layers of decision makers.
Gotta go even faster
It seems to be working, as the company was able to double its proposed AI orders last year. "We had set a target of a billion dollars in AI infrastructure product orders for the year, and we ended up doing over $2 billion, about double our original target," Patel said. In this week’s Q2 2026 earnings report, the company reported $2.1 billion in AI infrastructure bookings from hyperscalers, already exceeding its full-year 2025 total. That’s up from $800 million in Q4 2025 and $1.3 billion in Q1 2026 when the company first began recording these numbers.
Orders don't always translate into revenue. They still have to close the deals. Regardless, Patel is anything but satisfied. If anything, he wants the company to go faster than it has been and really seize the moment. "We're 95,000 employees, and my paranoia is always about, are we moving fast enough and getting enough done," he said.
If that's the case, it's a good problem to have. While Mueller likes Cisco's chances, he sees the company's AI revenue still representing only a small fraction of the $56 billion and change the company made last year, and he thinks it's still going to face an uphill battle turning AI from a niche driver into a meaningful part of the business.
Even so, Patel says at the end of the day no other company can offer the range of products and services that Cisco is producing, especially one that is combining networking, security and data in a single stack. "No other company has the combination of what we have," he said. Whether that will be enough remains to be seen.
