For many years, investors labeled Cisco as “stable, predictable, slightly dull.” A San Jose-based networking behemoth that rarely shocked anyone, paid dividends, and repurchased its stock. Then, half of the market was consumed by the AI trade, and for some reason, Cisco ended up in the middle of it. The story practically wrote itself in reverse, with the numbers coming first and the narrative catching up later. It has been strange to watch this play out.
On May 13, the third-quarter results were released, and they were truly impressive. With GAAP earnings per share of $0.85, a 37% increase, revenue reached a record $15.8 billion, up 12% year over year. Revenue from networking increased by 25%. The management increased its full-year AI infrastructure target from $5 billion to $9 billion after total product orders increased by 35%. This type of revision typically indicates either wishful accounting or actual demand. It seems to be the former in this instance, but it’s still unclear how long the order surge will last after the hyperscaler buildout slows.
The beat isn’t what makes the quarter intriguing. It’s the paradox that sits next to it. Cisco announced that it would lay off less than 4,000 employees, or less than 5% of its workforce, at a restructuring cost of up to $1 billion on the same day it reported record revenue. The phrase “really not a savings-driven restructure,” which makes you read it twice, was carefully chosen by CFO Mark Patterson. According to the company, it is pushing people into silicon, optics, security, and artificial intelligence instead of slower corners. In the tech industry, it’s becoming standard practice to post a strong number and still reduce the team. Depending on the earnings call you’re listening to, that may indicate discipline or anxiety.
The CEO and chair of Cisco, Chuck Robbins, has been living in the present. During his speech at JPMorgan’s technology conference, he referred to the current state of affairs as a “networking super cycle,” likening it to the dot-com era but with greater speed and scale. It’s important to note the self-awareness in that comparison because Cisco was a prime example of the 1999 boom and subsequent crash. This time, Robbins made a distinction: unlike many of the clients driving the internet boom in the early 2000s, the businesses investing heavily in AI now have solid balance sheets and actual cash flow. It’s a valid point. Additionally, that’s precisely what someone who is selling into a boom would say.
A 2016 acquisition that most people overlooked gave rise to Silicon One, Cisco’s in-house chip family, which serves as the story’s strategic backbone. Without its own silicon, Robbins stated bluntly that the $9 billion AI figure “would probably be close to zero.” Last quarter, Cisco announced five hyperscaler design wins, two of which were in optics. There’s a feeling that Cisco, having famously “missed the cloud revolution” (in Robbins’ own words), is determined not to miss this one, even if it means selling silicon-only to clients who previously thought of the company only as a hardware supplier.
Not everything is tidy. Revenue from services fell by 1%. There was no security. Tariffs contributed to the compression of gross margins. Additionally, Robbins’ constant reference to the campus refresh cycle—he used the baseball metaphor “top of the first”—remains more promise than evidence. Predictably, Reddit’s trading community posted six-figure gains and the customary “let’s party” threads, treating the print like nostalgia for 1999.
Beneath all of this, there is a slight tension that is difficult to ignore. By selling the plumbing beneath everyone else’s aspirations, Cisco has become an AI story by association, riding demand that it did not create. It’s a respectable spot to stand. Simply put, it’s not the same as being the reason someone came.
