Cisco reports record revenue, raises AI infrastructure forecast to $9B

Cisco reported record Q3 revenue of $15.8 billion, up 12% year over year, with adjusted EPS of $1.06 and net income of $3.4 billion. Product orders jumped 35%, including more than 50% growth in networking orders, and the company raised its full-year AI-related revenue forecast to $4 billion and AI infrastructure order outlook to $9 billion. Cisco also lifted fiscal 2026 revenue guidance to $62.8 billion-$63.0 billion, reinforcing strong demand tied to AI infrastructure buildouts.

Analysis

Cisco is increasingly functioning as a leverage proxy on hyperscaler capex, but the second-order winner is the broader networking ecosystem: the spend is migrating from generic enterprise refresh into AI-optimized switching, optics, and silicon-adjacent infrastructure where pricing power and mix can expand faster than headline revenue suggests. That favors the highest-basket-share suppliers with exposure to high-speed interconnect and campus refresh, while legacy box vendors with weaker software attach rates are at risk of being commoditized as buyers prioritize deployment speed and power efficiency.

The real incremental signal is not the size of the quarter, but the implied durability of order visibility into the next several quarters. If AI infrastructure orders are now a multi-year buildout rather than a one-off catch-up, the market likely underestimates how much operating margin can re-rate as utilization rises and restructuring costs wash through. The job cuts also imply management is using this cycle to rebase the cost structure around AI/security, which can create a cleaner earnings step-up in FY26 if demand holds.

The main risk is that this becomes a classic “sell the news” setup if hyperscaler capex slows after a few quarters of aggressive buildout or if customers rotate spending from networking into compute, where returns are easier to show. A second risk is execution: networking growth is only valuable if Cisco can sustain delivery, avoid supply bottlenecks, and preserve margin while ramping AI-related products. Any hint of order pull-forward, especially in the next 1-2 earnings prints, would compress multiple expansion quickly.

Consensus is likely still treating Cisco as a mature infrastructure compounder, but the move may be underowned because investors are anchoring on legacy enterprise exposure rather than AI network intensity. That creates room for a continued multiple rerate if AI orders remain above guidance, since networking is one of the few AI capex buckets where demand can stay strong even if model training economics get pressured. The trade is less about near-term EPS beats and more about whether Cisco can prove it is embedded in the AI buildout stack for 12-24 months, not just one quarter.

Key Decisions for Investors

• Long CSCO on pullbacks over the next 1-3 weeks; target a 10-15% upside rerating if AI order momentum persists through the next print, with downside limited by a still-reasonable earnings base and improving mix.
• Pair trade: long CSCO / short a slower-growth legacy networking peer basket over 1-2 quarters; thesis is that AI-linked demand and restructuring can widen relative growth/margin dispersion by 300-500 bps.
• Buy CSCO Jan-2027 calls or call spreads to express a 12-18 month AI-infrastructure rerating; risk/reward improves if the market starts capitalizing Cisco as a multi-year AI networking beneficiary rather than a mature hardware name.
• For more tactical exposure, fade any post-earnings weakness only if order growth decelerates sharply; otherwise hold through volatility, as the next catalyst is likely guidance credibility rather than the current quarter.