Cisco shares fell after hours despite a clean earnings beat, with investors zeroing in on a softer full-year outlook that undercuts the company’s AI-fueled narrative.
Cisco posted fiscal Q2 revenue of 15.35 billion dollars, up 10 percent year over year and ahead of the 15.12 billion dollar consensus. Adjusted earnings per share landed at 1.04 dollars, topping expectations of 1.02 dollars. The stock still dropped about 5.5 percent in late trading as full-year guidance missed the mark: management projected EPS of 3.00 to 3.08 dollars on revenue of 61.2 to 61.7 billion dollars, below Wall Street’s 3.12 dollars and 62.1 billion dollars, respectively. The message was blunt. Strong execution today is not enough if the growth curve flattens tomorrow. With the shares recently around the mid 80s, investors wanted a raise, not a reiteration and a haircut. They did not get it.
In a market primed for AI-led operating leverage, the bar is higher for blue chips positioned as infrastructure winners. Cisco cleared the quarter cleanly but failed to convince on the runway. A 10 percent top-line gain into a guidance range that implies slower momentum does not square with hopes for an AI capex supercycle spilling broadly across networking. The reaction suggests the Street is discounting a digestion period in enterprise and service-provider spending after last year’s restocking and elongated lead times. When a company leans on an AI story, the market expects that to translate into a stepping function in orders. Anything that looks like normalization or plateauing invites multiple compression.
Cisco’s Silicon One G300 arrived with hallmarks of a serious swing at the 600 billion dollar AI infrastructure market. Built on a 3-nanometer process, the chip is pitched to accelerate AI compute by 28 percent and cut energy use by 70 percent in liquid-cooled systems. That is the kind of silicon narrative investors want to hear. The problem is timing and scale. Even compelling specs need design wins, qualification, and deployment cycles before they show up in revenue. For now, the company still skews to switching, routing, and enterprise networking, while the heaviest AI wallet share remains concentrated with established acceleration and networking suppliers tied directly to leading AI platforms. The new chip can change Cisco’s trajectory over time, but the full-year guide says it will not change it fast enough for this fiscal year.
The selloff was as much about multiple as it was about models. On guided EPS of roughly 3 dollars, a stock price in the mid 80s implies a price-to-earnings multiple approaching the high 20s, rich against Cisco’s historical range and elevated for a company signaling more measured growth. That tension is showing up in research calls. New Street cut the stock to Neutral and lowered its target to 70 dollars, flagging a valuation that leaves less room for upside if execution is merely solid. When the path to earnings acceleration is cloudy and the multiple is already full, guidance misses sting harder. Cisco’s beat-and-lower dynamic invited that repricing in real time.
Softness in the outlook revives a familiar worry: that last year’s order recovery reflected channel restocking and elongated delivery schedules more than sustainable demand. HSBC trimmed its stance to Hold and warned that networking growth is not reaccelerating as hoped. As analyst Stephen Bersey put it, we expected Cisco’s networking segment to report improved growth, yet the company’s FY26 revenue guidance suggests the restocking effect may be coming to an end sooner than expected. That message dovetails with enterprise CIO caution this calendar year, especially in classic campus and branch refreshes. Cloud and telco spending remains uneven, with hyperscalers still prioritizing AI training clusters over general-purpose network upgrades. Cisco needs that second engine to kick in.
Cisco is not alone in chasing AI adjacency. The current capex cycle is dominated by a few vendors supplying accelerators, high-bandwidth memory, and specialized interconnects. Downstream beneficiaries in switching and optics have rallied when orders inflect and faded when deployment lags. For Cisco, the competitive set in data center networking is unforgiving, and share shifts tend to show up abruptly in orders, not press releases. With a major AI chip now in the bag, the company still needs to prove socket wins and volume shipments. Until investors can map those wins to a revenue bridge, the core portfolio’s trajectory carries more weight in the model—and this quarter’s guidance says to temper expectations.
Cisco has levers. A clearer inflection in product orders, particularly in high-speed data center switching tied to AI fabrics, would matter more than any slide deck. Evidence that Silicon One G300 is winning deployments at scale—paired with faster conversion from pilot to production—would help rerate the multiple. So would stability in campus and enterprise deals as backlogs normalize, plus margin resilience as mix shifts. The company’s services and software subscriptions can also buffer cyclicality if renewal rates hold and upsell improves. In a market that is paying up for durable growth, Cisco needs to turn AI promise into predictable bookings and show that the restocking unwind is not a cliff.
The tape told a story of frustration. Shares swung in heavy volume, with an intraday range that underscores how little patience remains for any wobble in outlooks from mega-cap incumbents. At roughly the mid 80s, the stock now prices a tighter spread between execution risk and AI optionality. That can work if orders reaccelerate or the company nudges guidance higher later this year. It can also compress quickly if enterprise budgets stall into midyear. For now, the market’s message is simple: beats are table stakes, guidance is the catalyst. Cisco cleared the first and missed the second.
Cisco delivered a better quarter and a weaker year. The earnings beat shows the operating machine is intact; the guide says growth is leveling as restocking fades and AI contributions remain nascent. With valuation already leaning ambitious for a networking incumbent, the burden of proof has shifted back to orders, backlog conversion, and evidence that the AI chip push is more than a headline. If those arrive, the multiple can hold. If not, the stock will trade the outlook, not the quarter.
