Cisco's AI Infrastructure Orders Hit $9.3 Billion As Networking Demand Surges

Cisco spent fiscal 2026 turning its AI infrastructure opportunity into a measurable business. Hyperscalers placed $9.3 billion in AI infrastructure orders with the networking company during the year, including $4 billion in the fourth quarter. Cisco expects approximately $7.5 billion in related revenue in fiscal 2027. 

Cisco's networking revenue reached $34.7 billion for fiscal 2026, up 22% from the previous year, while networking product orders rose 40% in the fourth quarter. Total product orders increased 35%. The spending behind those orders is reaching the physical infrastructure required to run AI workloads, from compute and networking to connectivity. AI spending is moving beyond software.

The Orders Came Faster Than Cisco Expected

Cisco had initially expected about $5 billion in hyperscaler AI infrastructure orders for fiscal 2026. The company reported $1.3 billion in the first quarter and $2.1 billion in the second. By the end of the third quarter, the total had reached $5.3 billion, prompting Cisco to raise its full-year expectation to $9 billion. 

Cisco took another $4 billion of AI infrastructure orders in the fourth quarter, bringing the fiscal-year total to $9.3 billion. It recognized approximately $4 billion in AI infrastructure revenue during fiscal 2026 and expects $7.5 billion in fiscal 2027. 

The company is now forecasting revenue from the business rather than setting a new annual AI infrastructure order target. The $7.5 billion figure will be one of the numbers investors watch as Cisco moves into its new fiscal year.

Cisco's fiscal 2026 revenue was $63.3 billion, up 12% from the previous year. 

What Cisco Is Selling Into The AI Buildout

Cisco is selling the networking equipment that connects the compute in an AI data center.

That requires networks capable of handling enormous volumes of traffic between processors, storage and other systems. Cisco has spent years developing its Silicon One networking architecture for that market.

In February, the company introduced the Silicon One G300, a 102.4-terabit-per-second switch chip aimed at high-bandwidth AI workloads. Cisco said its testing showed the platform could improve network utilization by 33% and reduce job completion time by 28%. 

Cisco is also building for AI workloads that stretch across data centers. Its Silicon One P200 powers the Cisco 8223 and N9000 systems, which Cisco is positioning for AI networking across multiple data-center locations. Cisco's Silicon One approach to AI infrastructure provides more detail on the technology.

Hyperscalers accounted for the $4 billion of AI infrastructure orders Cisco received in the fourth quarter. The customer base is not limited to the largest cloud companies.

In June, Cisco and NetApp expanded their FlexPod partnership with validated architectures for enterprise AI deployments, including AI inference, retrieval-augmented generation and edge computing. 

Strong Numbers, Higher Expectations

Cisco's fourth-quarter results cleared Wall Street's expectations. Adjusted earnings were $1.22 per share, versus the $1.17 expected by analysts, while revenue came to $17.25 billion against a $16.82 billion estimate. Revenue rose 17.6% year over year. 

The company forecast fiscal 2027 revenue of $72.2 billion to $73.4 billion, above the $68.69 billion analyst average compiled by LSEG. Cisco also expects first-quarter revenue of $18 billion to $18.2 billion. 

Shares fell more than 4% in after-hours trading following the August 12 report. Cisco's stock had already gained more than 60% in 2026 as investors anticipated stronger AI-related growth. 

Margins offered one reason for the caution. Cisco reported fourth-quarter non-GAAP gross margin of 66.3% and expects 65% to 66% in the first quarter. Reuters reported that the outlook was slightly below the market estimate, citing a more hardware-intensive product mix and higher component costs. 

Cisco has already shown that hyperscalers are willing to spend billions on its networking equipment. The next set of results will show how much of that spending turns into revenue, and what it costs the company to deliver it.

Brookfield Wealth Solutions estimates that the global AI buildout could require roughly $7 trillion in capital over the next decade across power generation, compute infrastructure and connectivity.

Cisco is selling networking equipment into that buildout. Fiscal 2027 will show how much of the $9.3 billion order haul turns into revenue, and what it costs the company to deliver it.