Cisco’s AI Orders Show Networking Is Now Part Of The Compute Boom

Cisco’s latest results show that the AI infrastructure boom is no longer only about chips, servers and data centres. Networking is now firmly inside the same spending cycle.

The company reported fourth-quarter and fiscal 2026 results on Wednesday, with Q4 revenue of $17.3 billion, up 18 percent year-on-year. GAAP net income came in at $3.9 billion, or $0.97 per share, while non-GAAP net income was $4.9 billion, or $1.22 per share.

The bigger story is guidance. Cisco expects fiscal 2027 revenue between $72.2 billion and $73.4 billion, ahead of Wall Street expectations cited by Reuters and other market trackers. It also expects non-GAAP earnings per share of $5.05 to $5.11 for the year. That is not the kind of forecast a mature networking company gives if AI demand is only a passing boost.

Cisco said it received $4 billion in AI infrastructure orders from hyperscale customers in the fourth quarter alone, taking total fiscal 2026 AI infrastructure orders to $9.3 billion. The company also said it delivered about $4 billion of AI infrastructure revenue during the year and expects roughly $7.5 billion in AI infrastructure revenue in fiscal 2027.

That matters because AI workloads need more than GPUs. Large training and inference clusters need high-speed networking, optics, switching, routing, security and observability. If compute is becoming the factory floor for AI, networking is the transport system that keeps the factory moving. Cisco wants investors to see it as part of that buildout, not as an old enterprise hardware vendor watching Nvidia take the spotlight.

The company’s networking business reflected that shift. Cisco reported product revenue of $13.1 billion in Q4, up 24 percent year-on-year, while networking revenue rose sharply. Demand from hyperscalers and AI data-centre customers is now giving Cisco a growth story that looks different from its traditional enterprise refresh cycles.

There are still reasons for caution. AI hardware supply chains are expensive, and Cisco’s gross margin has come under pressure as the mix shifts toward more AI infrastructure hardware. Investors also went into the report with high expectations after a strong year for the stock, so even upbeat results can meet a harsher market reaction if margins or forward commentary raise questions.

The strategic point is clear, though. We recently wrote that Nvidia is helping turn AI compute into a Wall Street asset class, and that Foxconn’s AI server boom is becoming bigger than its Apple story. Cisco’s numbers add the networking layer to that same picture.

For enterprise customers, this also signals where budgets are moving. AI projects that start as model experiments eventually require full-stack infrastructure: compute, networking, storage, security, power and software. Cisco’s advantage is that it already sells into the networks enterprises and cloud providers depend on.

The challenge is proving that AI orders become durable revenue rather than a temporary spending wave. Cisco has been through many technology cycles before, and investors know that hardware demand can be uneven. But for now, AI is giving the company a stronger growth argument than it has had in years. The routers and switches may not sound as glamorous as frontier models, but without them, the AI boom does not scale.