Cisco's Record Q4: $9.3B in AI Orders Ignites a Networking Supercycle

For years, Cisco was the company that built the routers and switches quietly humming in enterprise closets — essential infrastructure, but hardly the star of the AI revolution. That narrative changed dramatically on August 12, 2026, when the networking giant reported fourth-quarter fiscal year 2026 results that put it squarely at the center of the AI buildout.

Cisco posted record Q4 revenue of $17.3 billion, up 18% year-over-year, and full-year FY2026 revenue of $63.3 billion, up 12%. The company also issued forward guidance for fiscal 2027 that obliterated Wall Street expectations, projecting revenue between $72.2 billion and $73.4 billion — far above the analyst consensus of $68.69 billion. The message was unmistakable: the AI infrastructure gold rush is pouring directly into Cisco’s coffers.

The Numbers Behind the Beat

The headline metrics from Cisco’s Q4 and full-year FY2026 paint a picture of a company firing on all cylinders:

• Q4 revenue: $17.3 billion (+18% YoY)
• Full-year FY2026 revenue: $63.3 billion (+12% YoY)
• Non-GAAP EPS: $1.22 in Q4, beating the $1.17 consensus by 4.3%
• GAAP EPS: Up 52% year-over-year in Q4; full-year GAAP EPS of $3.33, up 31%
• Product revenue: Up 24% in Q4, with product orders surging 40% year-over-year
• Networking revenue: Up 28%, driven by data center demand

This marked the fifth consecutive quarter of EPS beats for Cisco, a streak that has fundamentally reshaped how investors view the company. Where Cisco was once seen as a mature, slow-growing networking vendor, it is now increasingly positioned as a critical AI infrastructure player.

AI Orders: The 4.5x Surge

The single most striking figure from the earnings report was the growth in AI infrastructure orders. Cisco booked $4 billion in AI-related orders in Q4 alone, bringing the full fiscal year 2026 total to $9.3 billion — approximately 4.5 times the fiscal year 2025 total.

The growth was not concentrated in a single customer segment. According to the earnings slides, Cisco saw triple-digit year-over-year increases across three key categories: neocloud providers, sovereign cloud deployments, and enterprise AI orders. In total, $1.3 billion in AI orders came from these diversified segments, signaling that demand has broadened well beyond the initial hyperscaler wave.

This matters because it suggests the AI infrastructure buildout is entering a more mature, distributed phase. It is no longer just a handful of cloud giants buying GPUs and networking gear by the truckload — governments, specialized cloud providers, and large enterprises are all now placing significant orders.

The Networking Supercycle Thesis

Cisco’s earnings slides explicitly framed the current environment as an “AI Unlocking a Cisco Networking Supercycle.” The logic is straightforward: AI training and inference workloads demand radically different network architectures than traditional enterprise IT. Massive GPU clusters require ultra-low-latency, high-bandwidth interconnects, and the networking gear that connects thousands of accelerators together has become a bottleneck as much as the silicon itself.

At the heart of Cisco’s AI networking strategy is its Silicon One platform, particularly the G200 processor — a 51.2 Tbps full-duplex switching processor designed for the massive-scale fabrics needed in AI data centers. Cisco has been integrating Silicon One with NVIDIA’s Spectrum-X Ethernet networking platform, creating a combined offering that positions both companies to capture the networking spend that accompanies every GPU deployment.

CEO Chuck Robbins has committed to migrating all of Cisco’s high-performance networking systems to Silicon One by fiscal year 2029, a transition that industry analysts at Hyperframe Research estimate could make Silicon One an $8-12 billion business on its own.

Splunk Integration Paying Off

The acquisition of Splunk, completed in early 2024 for $28 billion, is also beginning to pay dividends. On the Q3 earnings call, Cisco’s leadership indicated they were on track to exceed their target of 1,000 new customer logos for Splunk. The observability and security platform is increasingly being woven into Cisco’s broader AI strategy — the Secure AI Factory initiative leverages Splunk’s observability capabilities to monitor AI agent performance and data center health.

Security revenue, which had been relatively flat in recent quarters, is expected to benefit from deeper Splunk integration as Cisco bundles networking, security, and observability into unified AI infrastructure packages.

FY2027 Guidance: A Bold Bet

Cisco’s fiscal 2027 revenue guidance of $72.2-$73.4 billion represents growth of roughly 14-16% over FY2026 — a striking acceleration for a company of Cisco’s size. For Q1 FY2027, the company guided to non-GAAP EPS of $1.32-$1.34 and GAAP EPS of $1.08-$1.10.

The guidance implies that Cisco expects AI infrastructure spending to continue accelerating rather than plateauing. This aligns with the broader market signal from CoreWeave’s recent $104 billion backlog and NVIDIA’s continued supply-constrained demand. The AI capex cycle, far from showing signs of cooling, appears to be deepening and broadening.

The Paradox: Record Results, Falling Stock

Despite beating on both revenue and earnings, Cisco’s shares fell approximately 5% in after-hours trading following the report. This paradox reflects the high expectations that had been baked into the stock — Cisco’s shares had risen significantly in the weeks leading up to earnings as investors positioned for a blowout quarter.

The pullback also reflects lingering concerns about sustainability. Some analysts question whether the triple-digit growth rates in AI orders can persist, or whether Cisco is simply pulling forward demand that would have materialized over a longer period. Others worry about the competitive landscape, particularly from white-box networking vendors and NVIDIA’s own networking ambitions following its Mellanox acquisition.

However, the counterargument is equally compelling: AI infrastructure deployments are still in their early innings. The vast majority of enterprise AI workloads have not yet been deployed at scale, and the networking requirements for distributed inference — which many believe will dwarf training workloads — have barely begun to materialize.

The Restructuring: Funding the AI Pivot

In May 2026, alongside its Q3 earnings, Cisco announced a restructuring plan that would eliminate approximately 4,000 jobs — roughly 5% of its workforce. The cuts were explicitly framed as a reallocation of resources toward AI infrastructure, security, and the technologies driving the current growth wave.

This is a familiar pattern in the tech industry: companies that recognize platform shifts early and aggressively reallocate capital tend to emerge stronger. Cisco’s willingness to cut costs in legacy areas while doubling down on AI networking silicon and infrastructure suggests management understands the stakes.

What This Means for the AI Ecosystem

Cisco’s results are a bellwether for the broader AI infrastructure market. When the company that essentially invented enterprise networking reports that AI orders grew 4.5x year-over-year and guides to 14-16% revenue growth, it confirms that the AI buildout is not a bubble confined to GPU manufacturers — it is a systemic transformation of the entire networking and computing stack.

For investors, the key question is whether Cisco can maintain its competitive position as the networking layer becomes increasingly strategic. With Silicon One, the NVIDIA partnership, and Splunk’s observability capabilities, Cisco has assembled a compelling stack. Whether it can execute on that vision at scale — and whether the AI spending cycle has lasting power — will define the company’s trajectory for years to come.

One thing is certain: the era of Cisco being dismissed as a boring networking vendor is over. The AI supercycle has arrived, and Cisco is riding it.