Cisco's AI Hardware Windfall Tests Profit Discipline as Margins Tighten

Cisco's Q3 revenue hits record $15.84B on AI hardware surge, but software slowdown and margin squeeze prompt lowered EPS forecast and 4,000 job cuts.

The networking giant's latest quarterly results painted a picture of explosive demand for artificial-intelligence infrastructure – but beneath the record $15.84 billion in revenue, a less comfortable story is unfolding. While AI orders are reshaping the company's growth trajectory, they are also squeezing the profit margins that investors have come to expect from the software-heavy side of the business.

Cisco's network revenue surged 25 percent to $8.82 billion in the fiscal third quarter, driven by a roughly 30 percent jump in hardware sales to hyperscale cloud operators. That segment now accounts for the bulk of the order book: management lifted its forecast for AI-related orders in fiscal 2026 to around $9 billion. Yet the software business – traditionally a source of higher-margin subscription revenue – expanded only in the low single digits, shifting the product mix sharply toward less profitable equipment.

Chief Financial Officer Mark Patterson described the dynamic bluntly, warning that large hardware shipments to hyperscalers represent a "different gross-margin play" than Cisco's core software offerings. For the full fiscal year 2026, the company now guides for adjusted earnings per share of between $4.27 and $4.29 – a target that reflects the margin pressure. "We want to take the AI growth, but not at any cost," Patterson signaled, underscoring the need for price discipline and cost control.

A parallel push into security and network resilience is meant to offset some of that hardware drag. Cisco, together with Nvidia, Okta and OpenAI, recently launched the "EnterpriseClaw" initiative, integrating its own defense technologies to secure autonomous AI agents in corporate environments. The effort builds on the capabilities of Splunk, the monitoring and analytics subsidiary Cisco acquired last year. Splunk's study "The $600 Billion Wake-up Call" estimates that unplanned system outages now cost large global enterprises roughly $600 billion annually – a figure that has jumped 50 percent over two years. The median outage cost for the biggest firms stands at $15,000 per minute, translating into an average revenue loss of $95 million per company per year. Those numbers, Cisco argues, make resilience and security solutions a strategic imperative rather than a niche offering.

To refocus the organization on higher-growth areas such as silicon, optics, security and AI, the company is cutting about 4,000 jobs – fewer than 5 percent of its workforce. Chief Executive Chuck Robbins calls the current environment a "networking supercycle" and insists the restructuring is not a blanket cost-cutting exercise. Still, the transformation carries a price: restructuring charges could reach as much as $1 billion through fiscal 2027. Meanwhile, overall product orders climbed 35 percent, though collaboration revenue slipped, confirming the shift within the portfolio.

The stock market has taken note of the tension. After a furious rally that left Cisco up roughly 56 percent year-to-date as one of the most overbought names in the S&P 500, the shares closed at $99.42 (approximately €99.42) on Tuesday – down from the day's earlier level and off the 52-week high. The relative strength index (RSI) sits at 76, still deep in overbought territory. Adding to the caution, executives Deborah Stahlkopf and Jeetendra Patel sold shares in mid-May under prearranged trading plans.

Despite the near-term headwinds, analysts remain broadly bullish. HSBC upgraded the stock to Buy with a $137 price target, while UBS and Rosenblatt set targets as high as $150. Argus followed suit with a $150 target, seeing roughly 30 percent upside from current levels. For income-oriented investors, the next payout is scheduled for July – a quarterly dividend of $0.42 per share.

The challenge for Cisco going forward is balancing the undeniable momentum from AI infrastructure spending with the profitability discipline that the market now demands. Every incremental hardware order from a hyperscaler brings revenue – but also underscores the urgency of expanding software attach rates and maintaining pricing power in a competitive landscape.