Cisco Posts a Record $18.5 Billion Quarter as AI Infrastructure Orders Surge

Cisco spent years as the steady, almost sleepy name in tech earnings season. That script didn’t survive November 12, 2026. Reporting its first-quarter results for fiscal 2027, Cisco posted record revenue of $18.5 billion, up 24% from the same quarter a year earlier, with GAAP earnings per share of $1.14, up 58%. Those aren’t my adjectives, either. “Record” is the word Cisco itself used, and this time the numbers genuinely back it up.

The headline figures only tell half the story, though. Management also raised the company’s full-year outlook, and said it now expects roughly $9 billion in fiscal-year AI infrastructure revenue from hyperscale customers alone. So what’s really going on here? A networking company is starting to look like one of the clearest scoreboards for the AI buildout, and this quarter is the strongest evidence yet.

Record Numbers, Reported Straight From Cisco

Let’s start with what the company actually reported, because the framing matters. Revenue of $18.5 billion and GAAP earnings per share of $1.14 are Cisco’s own reported results for the quarter, published through Cisco Investor Relations on November 12, 2026. A 24% revenue jump is the kind of growth rate you’d expect from a young cloud startup, not from a networking giant that has been selling routers and switches for decades.

The guidance raise tells you how management reads its own order book. Cisco now expects fiscal 2027 revenue in the range of $74.8 billion to $75.8 billion, a step up from where it had pointed investors before. Companies don’t lift full-year guidance in the first quarter unless the orders already in hand give them real confidence. Though confidence and delivery are two different things, and we’ll come back to that.

It’s also worth remembering who Cisco sells to. This isn’t a consumer brand riding a viral product cycle. Its customers are enterprises, telecom operators and the largest cloud providers on the planet, and those buyers plan spending quarters in advance. When Cisco’s revenue moves this sharply, it usually means a lot of separate boardrooms reached the same conclusion at roughly the same time: spend on the network, now.

The $9 Billion Hyperscaler Signal

The number that matters most for the wider AI industry isn’t the revenue. It’s the expectation. Cisco says AI infrastructure revenue from hyperscale customers should reach about $9 billion across its fiscal year. Hyperscalers are the giant cloud operators building the enormous data centres that train and run AI models, and they are spending at a pace that keeps pulling their suppliers along with them.

You can see that spending from the other side, too. Google recently bought Crusoe’s unfinished Wyoming data centre site as part of a long-term investment in the state, a deal we covered in our report on Google’s Crusoe Wyoming data centre move. Every facility like that needs far more than processors. It needs thousands of fast, reliable connections between servers, and that connective layer is exactly what Cisco sells.

The chip announcements grab the headlines, of course. Google’s eighth-generation accelerators, which we looked at in our piece on the Google TPU 8tx and TPU 8i Max, are built for agentic AI workloads that hammer the network constantly. But accelerators only earn their keep once they’re wired into clusters that can move data fast enough to keep them fed. Cisco’s $9 billion expectation is, in effect, a measurement of how much wiring the AI era still needs.

Why Networking Keeps Winning the AI Spending Race

There’s a simple reason networking vendors keep showing up in AI earnings stories. Training a large model isn’t one computer doing one job. It’s thousands of processors swapping data back and forth without pause, and if the network between them slows down, very expensive hardware sits idle. The faster the chips get, the more brutal that bottleneck becomes. So each new generation of AI hardware quietly creates a matching order for switches, optics and routing gear.

Cisco is also catching a second wave: the enterprise refresh. Plenty of companies delayed network upgrades while budgets went to cloud bills and, lately, to AI experiments. Those ageing office networks now have to carry AI traffic they were never designed for. When a business wires itself up for AI tools, the order rarely stops at software licences. New switches and wireless gear tend to follow close behind.

Put the two together, hyperscale data centres on one side and enterprise refreshes on the other, and a 24% revenue jump stops looking like a fluke. It looks like a company standing in the path of two spending waves at once. Cisco’s own updates on the Cisco Newsroom have been hammering the AI infrastructure theme for months; this quarter, the financial results finally matched the messaging.

Raised Guidance Is a Promise, Not a Result

Now the fine print, because it matters. Everything in the outlook, including the $9 billion AI infrastructure expectation, is a forward-looking statement from Cisco’s management. It describes what the company expects, based on the orders and demand it can see today. It is not a guarantee, and Cisco’s release carries the standard legal language saying exactly that.

Hyperscaler spending can shift, too. A cloud giant that delays one data centre phase can move a supplier’s quarter. Power availability, construction timelines and borrowing costs all sit between an order and installed, revenue-earning equipment. And a raised outlook cuts both ways: it lifts expectations, which means every future quarter now gets judged against a higher bar.

Still, I keep coming back to the direction of travel. Not long ago, the debate around Cisco was whether it would miss the AI era entirely. This quarter, the company is reporting record revenue, raising its outlook and putting a $9 billion figure on hyperscaler AI demand. Whether the orders keep landing at this pace is the question that will decide if this was a peak or a pattern.