Cisco Systems shares surged approximately 17 percent in Thursday trading after the company reported record quarterly revenue of $15.8 billion on Wednesday evening and then stunned analysts with guidance that was so far above expectations it forced an immediate and significant repricing of what the stock is worth.
The move puts Cisco near $119, up from a close of $101.87, and represents the largest single-day gain for the networking giant in years.
The number that sent the stock higher was not the revenue beat, though that was meaningful.
It was the AI infrastructure order guidance that Cisco raised from $5 billion for the full year to $9 billion, nearly doubling the target, after booking $5.3 billion year-to-date in just three quarters.
And it was the Q4 revenue guidance of $16.7 to $16.9 billion against analyst expectations of $15.82 billion.
A company does not guide revenue $900 million to $1.1 billion above consensus without something real happening in its business.
“Cisco is well-positioned as the critical infrastructure for the AI era,” CEO Chuck Robbins said. The market’s response Thursday morning is the investment community agreeing with him.
What The Earnings Actually Showed
Cisco reported fiscal third quarter revenue of $15.8 billion, a record for the company and a 12 percent increase from the $14.15 billion it reported in the same period a year ago.
GAAP net income was $3.4 billion or $0.85 per share, up from $2.49 billion or $0.62 per share a year ago.
Non-GAAP net income of $4.2 billion or $1.06 per share came in above the high end of prior guidance on both lines.
The segment breakdown tells the AI story precisely. Networking revenue, the business of selling the switches, routers and optical equipment that connects AI data centers, grew 25 percent year-over-year.
Product orders overall grew 35 percent. Networking orders specifically grew more than 50 percent.
Hyperscaler orders, the purchases from Amazon, Microsoft, Google and Meta building the massive AI computing infrastructure that defines the current capital expenditure cycle, grew at triple-digit rates in the quarter, with $1.9 billion booked in the single quarter alone.
CFO Mark Patterson confirmed that AI hyperscale revenue will be at least $6 billion in fiscal 2027, the first time the company has formally guided a number for the following fiscal year in this category, and that the Q4 guidance alone implies 14.5 percent top-line growth, a rate that is higher still than the 12 percent the company just reported for the record quarter.
The Number That Changed Everything
The guidance revision that deserves the most attention is the AI hyperscaler order target.
Entering the quarter, Cisco’s full-year FY2026 target for AI infrastructure orders from hyperscalers was $5 billion, and the company had already achieved $3.1 billion through Q2. Wall Street was watching to see whether the trajectory was accelerating or decelerating.
It was accelerating. The company booked $1.9 billion in Q3 alone, bringing the year-to-date total to $5.3 billion, already surpassing the entire prior full-year target with one quarter remaining.
Management promptly raised the FY2026 hyperscaler order target to $9 billion from $5 billion. That is not a minor revision.
That is a company telling investors that what was believed to be the ceiling was actually a floor, and that the actual ceiling is 80 percent higher.
The AI infrastructure revenue target for FY2026 was also raised, to $4 billion from $3 billion, representing a 33 percent increase in just one quarter’s update.
With Patterson’s $6 billion FY2027 floor guidance, Cisco has established a multi-year revenue ramp from AI infrastructure that analysts are now scrambling to incorporate into their models.
Cisco CEO Robbins noted that the growth is driven by design wins, meaning hyperscalers are specifically choosing Cisco’s Silicon One networking chip and Acacia optical products for their AI data center builds rather than alternatives.
Design wins create a specific kind of revenue durability, once a hyperscaler designs their data center around a particular networking architecture, switching costs are high and the relationship tends to persist across multiple generations of infrastructure.
The AI data center build-out that is driving Cisco’s orders today will generate maintenance, upgrade and expansion revenue for years.
What Cisco Does And Why AI Needs It
Cisco is not a company that most casual technology investors associate with the AI wave. Nvidia is the AI chip story.
Cerebras is the alternative chip story. Nebius and CoreWeave are the AI cloud infrastructure stories. Cisco is the company that makes the routers and switches.
That perception is the specific reason the stock was up only 30 percent year-to-date before Thursday, respectable performance but not the explosive re-rating that AI-adjacent companies have experienced.
The earnings report Thursday is the moment the market revised what it thinks Cisco actually is in the AI era.
The answer is that Cisco is the pipes. Every Nvidia GPU that sits in a data center needs to communicate with every other GPU to run the distributed AI training and inference workloads that large AI models require.
The faster and more reliably those GPUs can communicate, the more efficient the AI computation becomes.
Cisco’s networking equipment, specifically its Silicon One programmable networking chip and its Acacia high-speed optical interconnect technology, is what allows hyperscalers to build the data center architecture that makes GPU clusters function as a unified computing resource rather than a collection of isolated processors.
When Nvidia sells more GPUs to hyperscalers, Cisco sells more networking equipment to connect them.
The relationship is not incidental. It is structural. And the AI data center build-out that Nvidia has been benefiting from for the past two years has been flowing through Cisco’s networking revenue with a lag that the market was not fully pricing until Thursday’s report made the numbers undeniable.
The Restructuring Running Alongside The Record Quarter
Even as Cisco reported record revenue and raised guidance, CEO Robbins confirmed that the layoffs announced earlier this week are beginning today, May 14.
The restructuring will result in $1 billion in pre-tax charges, with approximately $450 million recognized immediately. Resources are being reallocated from lower-growth areas toward silicon, optics, security and AI.
Robbins framed the restructuring as a complement to the growth story rather than a contradiction of it. He said:
“The companies that will win in the AI era will be those with focus, urgency, and the discipline to continuously shift investment toward the areas where demand and long-term value creation are strongest. This means making hard decisions, about where we invest, how we’re organized, and how our cost structure reflects the opportunity in front of us.”
In a period when multiple large technology companies are cutting headcount while reporting strong revenue numbers, Cisco’s version of that combination is the most clearly explained.
The networking and AI infrastructure business is growing faster than the company’s cost structure anticipated.
The resources that were allocated to lower-growth areas are being redirected toward the areas where the $9 billion in AI orders and the triple-digit hyperscaler order growth are happening.
How Cisco Compares To Its Networking Peers
The 17 percent move for Cisco on Thursday put it significantly ahead of its closest comparables.
Arista Networks, the other major AI data center networking player, moved modestly but did not see comparable gains because the Cisco report raises questions about competitive dynamics in the hyperscaler AI networking market.
Arista has been the market’s preferred pure-play AI networking stock, and Cisco’s results suggest the AI data center opportunity is large enough and Cisco’s competitive position strong enough to support both companies while potentially narrowing the re-rating gap that had developed between them.
Hewlett Packard Enterprise, which serves a different mix of enterprise networking and AI infrastructure, underperformed Cisco on Thursday because its business does not have the same hyperscaler AI data center exposure.
The full-year FY2026 revenue guidance of $62.8 to $63 billion and the FY2027 AI hyperscale revenue floor of $6 billion give Cisco a forward narrative that is now competing for the same AI infrastructure investment thesis as companies that investors had previously viewed as more directly AI-exposed.
