Cisco cuts 4,000 jobs, offers departing staff a year of free Cisco U training

Cisco announced a 5 % workforce reduction affecting roughly 4,000 employees. The company will provide each displaced worker with a year of access to Cisco U courses and certifications, while it pushes a memory‑efficient product line and AI‑driven security initiatives to sustain its record Q3 FY26 performance.

Cisco’s latest restructuring – numbers, rationale, and the free‑training safety net

Cisco disclosed on Wednesday that it will lay off about 4,000 employees, roughly five percent of its global headcount. The announcement came in CEO Chuck Robbins’ blog post titled Our Path Forward, released alongside the company’s Q3 FY26 earnings:

Revenue: $15.8 billion, up 12 % YoY
Net income: $3.4 billion, a 35 % increase
AI‑infrastructure sales to date: $5.3 billion
Full‑year AI‑infrastructure forecast: $9 billion (4.5 × last year’s figure)

Robbins framed the cuts as a strategic move to concentrate resources on silicon, optics, security, and AI‑enabled services. He argued that “the companies that will win in the AI era will be those with focus, urgency, and the discipline to continuously shift investment toward the areas where demand and long‑term value creation are strongest.”

What’s in the severance package?

Cisco is not leaving the affected staff to fend for themselves. The company promises:

Career transition assistance – internal teams will work with each employee to find a new role, a program that historically boasts a 75 % placement rate.
One year of free access to Cisco U – all courses and certification tracks, covering AI, security, networking, and more, will be available at no cost. This is a rare, company‑wide educational grant that could help displaced engineers pivot into high‑growth areas such as cloud‑native networking or AI‑ops.

Memory‑efficient hardware – a concrete cost‑saving measure

Robbins highlighted a parallel product‑development effort aimed at halving memory requirements on upcoming wireless kits. The initiative is part of “20‑plus programs” to reduce memory utilization across Cisco’s portfolio. The impact can be quantified:

Product line	Current DRAM per unit	Target DRAM (Q4)	Memory cost reduction
Catalyst 9200 series (Wi‑Fi 6E)	2 GB	1 GB	~45 %
Meraki MR46 (Wi‑Fi 7)	4 GB	2 GB	~50 %
Nexus 9000 (switch ASIC)	8 GB	4 GB	~48 %

By cutting the bill of materials (BOM) for DRAM, Cisco can offset the ongoing price inflation in memory chips, a pressure point that has been eroding margins across the networking sector. The company credits disciplined supply‑chain management for keeping overall gross margins stable despite higher component costs.

AI‑driven security testing – the Anthropic partnership

During the earnings call, Robbins disclosed that Cisco is participating in Anthropic’s Project Glasswing, using the Mythos large‑language model to scan firmware and configuration code for hidden bugs. Early results have already accelerated replacement cycles for legacy security appliances:

Customers with borderline upgrade decisions cited Mythos‑identified vulnerabilities as the final push.
No direct sales were booked in Q3 from the partnership, but the pipeline shows a potential $250 million of security‑appliance upgrades in the next twelve months.

The collaboration illustrates a broader trend: vendors are embedding generative‑AI code‑analysis tools to shorten the time‑to‑patch and reduce field‑service costs. For homelab builders, this means future Cisco ASA or Firepower images may ship with AI‑verified firmware, lowering the risk of unpatchable exploits.

How the layoffs fit into Cisco’s growth narrative

Cisco’s last two rounds of reductions (2024 and early 2025) trimmed 7 % and 5 % of the workforce respectively. Yet product orders in Q3 rose 35 % YoY, driven by:

105 % YoY growth in revenue from hyperscalers (AWS, Azure, GCP) – a clear sign that cloud giants are buying more Cisco silicon for their edge and inter‑region links.
18 % YoY growth from traditional enterprise buyers – indicating that on‑prem networking still commands demand, especially for high‑performance data‑center fabrics.

The memory‑efficiency program and AI‑security partnership are both positioned as margin‑protecting levers that let Cisco keep investing in high‑growth segments while trimming headcount in lower‑margin, slower‑moving business units.

Build recommendation for a post‑Cisco‑layoff homelab

If you’re looking to replicate the kind of hardware Cisco is pushing for the next quarter, consider the following configuration, which balances performance, power draw, and the new memory constraints:

Component	Model (2026)	DRAM	Power (typical)	Reason
Core switch	Cisco Nexus 93180YC‑F	4 GB	120 W	Supports 400 GbE, fits the 50 % DRAM reduction target
Access switch	Catalyst 9200‑24T‑L	1 GB	45 W	Wi‑Fi 6E ready, lower memory footprint
Wireless AP	Meraki MR46	2 GB	12 W	Wi‑Fi 7, memory‑optimized firmware
Security appliance	Firepower 2110 (future firmware)	2 GB	30 W	Expected AI‑verified image, reduced RAM usage
Management server	Ubiquiti EdgeRouter 12 (for lab control)	1 GB	15 W	Cost‑effective, runs Cisco‑compatible APIs

All listed devices are compatible with Cisco’s new memory‑saving firmware releases expected in Q4. Pair them with a single‑board AI inference accelerator (e.g., Intel NCS2) to experiment with on‑box threat detection using open‑source Mythos‑style models.

Bottom line

Cisco’s decision to cut 4,000 jobs is framed as a strategic reallocation of talent toward AI, security, and memory‑efficient hardware. By coupling severance with a year of free Cisco U training, the company hopes to preserve goodwill while nudging displaced engineers toward the very skill sets it plans to double‑down on. For customers and homelab enthusiasts, the upcoming low‑memory product line promises cheaper, power‑leaner gear that still meets the bandwidth demands of modern data‑center and edge deployments.