Cisco Cut 5,600 Jobs While Profitable — What It Means for Tech Workers in 2026

Cisco CEO Chuck Robbins sat down with The Verge's Decoder podcast to talk about something a lot of tech workers have been trying to make sense of for two years: why a profitable, 80,000-person networking giant keeps cutting jobs while talking up its future. His answer matters because Cisco is one of the clearest bellwethers for where AI infrastructure money is actually going — and which tech roles are getting rebuilt around it. At Metaintro, we track these shifts so job seekers can move before the headlines do, and Robbins's interview is full of signals worth unpacking.

Why did Cisco lay off thousands of workers if business is good?

Robbins's framing on the Decoder interview was blunt: the layoffs were not about survival, they were about reallocation. Cisco has cut roughly 5,600 jobs across two major rounds since early 2024, totaling about 7% of its global workforce, according to filings tracked by Layoffs.fyi and reporting from Reuters. The company was profitable in every one of those quarters. So why the cuts?

The answer Robbins gave The Verge mirrors what he has told investors on earnings calls: Cisco needs to free up payroll dollars to hire aggressively in AI networking, security, and cloud infrastructure. In other words, the company is not shrinking — it is swapping. Engineers who built and maintained the legacy switching and routing business that defined Cisco for thirty years are being replaced, in budget terms, by engineers who can build silicon, optical interconnects, and AI fabric for hyperscale data centers.

That distinction matters for anyone reading the headlines and assuming "Cisco layoffs" means "tech is collapsing." It does not. It means Cisco is making a bet that the next decade of revenue comes from selling picks and shovels to companies building AI clusters — and the org chart it had in 2022 was not built for that bet. The painful part is that the people being cut are real, and many of them are mid-career engineers in roles that simply do not exist on the new map.

For tech workers, the takeaway is uncomfortable but useful: when a profitable company does layoffs, the explanation is almost never "we ran out of money." It is "we ran out of patience for the old roadmap." Reading press releases through that lens helps you predict where the next round will land — and where the new openings will open up.

Is the AI bubble real, and does Cisco think it will pop?

Robbins did not dodge the bubble question, and his answer was more interesting than the usual CEO talking points. He told The Verge that he believes the AI infrastructure buildout is durable, not speculative — but he was careful to separate two things: the spending on data centers, GPUs, and networking gear (which he called real and sustained) versus the sky-high valuations of some AI application companies (which he was less willing to defend).

His reasoning tracks with the numbers. Hyperscalers — Microsoft, Amazon, Google, and Meta — have collectively committed more than $300 billion in capital expenditure for 2025 and 2026, most of it earmarked for AI compute, power, and networking. NVIDIA's data center revenue has crossed $30 billion per quarter. Cisco's own AI infrastructure orders have grown into the billions, and the company has publicly committed to selling more than $1 billion in AI networking products to webscale customers.

If Robbins is right, the bubble debate is asking the wrong question. The risk is not that data center spending collapses — it is that the application layer (chatbots, copilots, agentic startups) takes longer to monetize than the infrastructure cycle assumed. That gap could squeeze the companies furthest from the metal first: AI startups with large compute bills and unclear revenue, then the consulting and integration shops that depend on those startups buying their services.

For job seekers, that is a meaningful map. Roles tied to physical AI infrastructure — networking, power, cooling, security, data center construction — sit on the safer side of the bubble line. Roles tied purely to the application layer carry more risk, especially at venture-funded startups that have not yet proven a paying customer base.

What is the memory shortage Robbins flagged, and why does it affect hiring?

One of the most underreported parts of the Decoder interview was Robbins's comment on a global memory and component shortage that is now stretching into 2026. He told The Verge that lead times on certain high-bandwidth memory and optical components have grown long enough that Cisco is having to plan its data center product launches around supply, not demand.

This matches what analysts at Gartner and IDC have been flagging since late 2025: HBM3 and HBM4 memory, used in every modern AI accelerator, is sold out through most of 2026. Optical transceivers, the unsexy components that move data between GPUs and racks, are also in tight supply. That bottleneck is reshaping hiring in two ways.

First, it is creating real demand for supply chain engineers, hardware program managers, and component sourcing specialists at every networking and infrastructure company — not just Cisco. If you can navigate semiconductor lead times, allocate constrained inventory, and negotiate with foundries, you are extremely employable right now. Job postings on LinkedIn for "supply chain engineer AI infrastructure" have more than doubled year over year.

Second, it is slowing some hiring cycles. When a product launch is gated by component availability, hiring for the engineering teams that support that product gets paced to match. This is part of why some AI infrastructure companies have hired in bursts rather than steady ramps — and why tech workers should pay attention to component announcements, not just funding rounds, when timing a job search.

What roles is Cisco actually hiring for in 2026?

Robbins spent a meaningful chunk of the interview talking about where Cisco is adding people, and the list lines up with the company's earnings commentary. The growth areas are AI networking (silicon, optical, switch fabric), security (Cisco's Splunk acquisition added thousands of engineers and the company is still hiring there), observability, and cloud-managed networking.

According to Cisco's careers site and recent filings tracked by BLS industry data, the company has thousands of open roles globally, weighted heavily toward engineering. Notable in-demand titles include AI networking software engineer, silicon design engineer, security threat researcher, data center solutions architect, and customer experience engineer for hyperscale accounts. Sales and go-to-market roles tied to AI infrastructure customers are also expanding.

What is not growing: legacy switch and router product lines, traditional collaboration (the Webex business has been restructured), and certain back-office functions that have been consolidated or outsourced. If your resume is anchored to those areas, the path forward is either internal reskilling — Cisco has been public about retraining programs — or a lateral move to a competitor or customer that still values that experience.

The broader lesson from Robbins's interview is that "tech hiring" is no longer one market. It is at least three: an AI infrastructure boom that is hiring as fast as it can find talent, a steady-state enterprise software market that is hiring cautiously, and a legacy hardware and services market that is contracting. Knowing which of those three you are in changes everything about how you write your resume, where you apply, and how you talk about your experience in interviews.

For tech workers caught in Cisco's recent rounds — or watching nervously from the sidelines at other networking and hardware companies — the most useful reframe from this interview may be the simplest one. Layoffs at profitable companies are almost always portfolio decisions, not panic. The job market that emerges on the other side rewards people who can show they understand the new portfolio — and who can speak the language of AI infrastructure even if they spent the last decade somewhere else.

People Also Asked

Q: How many people has Cisco laid off in total?

A: Cisco has cut roughly 5,600 jobs across two major rounds announced in February and August 2024, representing about 7% of its global workforce. The company has continued smaller, targeted reductions into 2025 and 2026 as it reshapes teams around AI infrastructure, security, and data center growth areas.

Q: Is Cisco still hiring in 2026?

A: Yes. Despite the layoffs, Cisco has thousands of open roles globally, concentrated in AI networking, silicon design, security (including its Splunk division), data center solutions architecture, and customer experience engineering for hyperscale accounts. The company is reallocating headcount, not shrinking overall.

Q: What did Chuck Robbins say about the AI bubble?

A: On The Verge's Decoder podcast, Robbins pushed back on the idea that AI infrastructure spending is a bubble, arguing the buildout is durable because hyperscaler capex is real and sustained. He was more cautious about valuations in the AI application layer, where monetization is still unproven for many startups.

Looking for your next opportunity? Metaintro tracks hiring at networking and AI infrastructure companies. Sign up free