Cisco Cuts 4,000 Jobs on the Day It Reports $15.8B Record Revenue—and the Paradox Is the Point

On May 14, Cisco Systems did something that would have been unthinkable at most companies: it announced record quarterly revenue of $15.8 billion—a 12% year-on-year gain—and on the same day notified approximately 4,000 employees that their jobs were being eliminated. The restructuring charge is expected to cost up to $1 billion in severance and related expenses. The stock rose.

This is the new math of the AI transition in enterprise technology, and Cisco is not the first company to run it. What makes the Cisco case worth examining is the specificity of where the company is cutting and where it is redirecting investment—a case study in how a $230 billion legacy technology incumbent is trying to navigate a fundamental shift in what its customers want from a network.

The Numbers Behind the Pivot

Cisco’s Q3 FY2026 revenue of $15.8 billion came in above analyst consensus, driven by two forces that are increasingly inseparable: hyperscaler spending on AI infrastructure and the networking requirements of AI data centers.

The company reported $5.3 billion in AI infrastructure orders year-to-date through the quarter’s close, and raised its full-year projection to approximately $9 billion. This represents a dramatic acceleration from the prior year, when AI-related networking orders were still a minority of enterprise infrastructure spending. The shift has been particularly pronounced in data center switching—the high-speed interconnects that link thousands of GPUs into coherent training clusters—and in optical networking equipment that moves data between facilities at speeds that traditional copper-based systems cannot approach.

Cisco’s Ethernet-based AI fabric products, which compete with Nvidia’s InfiniBand-dominated interconnect technology inside GPU clusters, have gained meaningful traction among hyperscalers seeking to avoid single-vendor dependency. Multiple cloud providers have disclosed plans to standardize portions of their AI infrastructure on Ethernet-over-optics configurations where Cisco hardware is the primary supplier, a positioning that gives the company a foothold in the AI compute layer it did not have two years ago.

What Is Being Cut—and Why

The 4,000 positions being eliminated—roughly 5% of Cisco’s global workforce of approximately 80,000—are concentrated in areas the company describes as “slower-growing businesses” and “legacy product lines.” These include teams supporting older enterprise routing platforms, on-premises telephony systems, and certain enterprise software segments that have struggled to grow in a market increasingly oriented toward cloud-native alternatives.

The pattern mirrors what other large technology companies have done over the past 18 months: Cloudflare, GitLab, Snap, and others have all announced layoffs concurrent with strategic pivots toward AI-native products. The common thread is not financial distress but reallocation—moving resources from businesses that are technically functioning but strategically marginal toward businesses that the company believes define the next competitive era.

For Cisco specifically, the strategic logic involves a difficult reality: its core enterprise routing and switching business, while still enormous, is growing at mid-single-digit rates at best in a market where hyperscaler AI-related infrastructure is growing at 40% to 60% annually. Companies that fail to align their cost structures and talent bases with faster-growing segments will, over time, find their competitive positions eroded even if their absolute revenues remain impressive.

The Human Cost of Efficiency

What makes the Cisco announcement particularly stark is the optics: eliminating 4,000 jobs on the same day as reporting record revenue removes any ambiguity about whether the layoffs are financially motivated. They are not. Cisco is profitable, cash-generative, and growing. The cuts are structural—a judgment that certain kinds of human expertise are less valuable in the AI era than they were before it, and that capital is better deployed in infrastructure, silicon development, and AI-native software capabilities.

This calculation is being made with increasing frequency across the technology sector, and it is generating real debate about what obligations profitable companies have to their workforces. Cisco’s chief human resources officer, in an internal communication viewed by multiple outlets, framed the restructuring as “investing in our people’s future by focusing them on the areas where Cisco will lead”—a formulation that has been met with skepticism by employees whose roles are being eliminated.

The severance terms, by most accounts, are generous—multiple months of compensation plus extended benefits—which is consistent with Cisco’s historical treatment of restructuring-affected employees. But the cumulative signal across the tech industry is unmistakable: AI is enabling companies to produce significantly more output per employee, and the organizations that are restructuring first may be the ones best positioned to compete in the next phase.

What Cisco Is Building Instead

The capital being redeployed through this restructuring is flowing in several directions. Cisco has significantly expanded its investment in Silicon One, its proprietary ASIC (application-specific integrated circuit) for high-performance networking, which now powers some of the fastest data center switches available and is winning deals with hyperscalers that were previously Arista or Juniper customers.

The company is also investing heavily in optical networking through Cisco Optics, building on the Acacia acquisition it completed in 2021. AI data centers require vast amounts of coherent optical transceivers to move data between GPU racks and between facilities; Cisco’s optical business is growing at a rate that outpaces the overall company and commands better margins than traditional switching hardware.

On the software side, Cisco’s AI networking management platform—which uses machine learning to autonomously optimize traffic flows across hybrid cloud environments—is being marketed to enterprises managing multi-cloud AI deployments that are too complex for traditional network operations center approaches.

A Template for AI-Era Incumbents

The Cisco story is, in microcosm, the story of how established technology companies are navigating the AI transition. The companies that articulate a credible path from legacy revenue to AI-era positioning—and execute on it with sufficient conviction to restructure simultaneously with growth—are being rewarded by markets. Those that continue to optimize legacy businesses without making visible structural bets on AI are accumulating strategic risk.

Whether Cisco’s specific bets—Ethernet AI fabric, Silicon One, coherent optics—prove to be the right ones will take years to determine. The hyperscaler orders suggest the market is paying attention. The $9 billion AI infrastructure order projection for FY2026, if it materializes, would represent roughly 15% of total revenue from a business line that barely existed three years ago.

The 4,000 employees receiving separation notices this week are, in a very direct sense, funding that trajectory.