Skip to main content
CybersecurityVulnerability Management

Cisco CEO Warns of Growing Risk from Unpatchable Technology

Rows of outdated servers and routers in a network operations center with technicians in the background.

"There’s a lot of concern about unpatched technology and their infrastructure, not just ours," Chuck Robbins told investors, arguing that AI tools are accelerating the push to modernize equipment that can no longer be patched.

Robbins frames Anthropic's Mythos as a security and modernization catalyst

Cisco’s chief executive described Anthropic’s Claude Mythos Preview as more than a productivity play inside the company: Robbins said Mythos is helping Cisco test its own code, accelerate patching and create incentives for customers to replace "last-day-of-support" equipment and other technology that "can’t be patched." He told investors the company has been embedding the technology into internal operations and that, while Cisco did not record any meaningful orders in the fiscal quarter ended April 25 as a result of Mythos, that could change as Cisco works with clients.

Cisco’s AI infrastructure demand and the company’s revised outlook

Robbins said Cisco has taken in $5.3 billion in AI infrastructure orders from hyperscalers over the past three quarters and now expects $9 billion for the full fiscal year — a figure the source describes as nearly double the company’s previous forecast. Those orders and investor expectations helped drive a strong earnings print: revenue of $15.84 billion for the quarter ended April 25 beat Seeking Alpha's sales estimate of $15.56 billion, and Cisco’s non‑GAAP earnings of $1.06 per share edged past Seeking Alpha's non‑GAAP estimate of $1.04.

The market response was immediate: Cisco stock rose $20.13 — or 19.76% — to $122 per share in after‑hours trading Tuesday, the highest intraday level reported since the company went public in February 1990. For the fiscal quarter ending July 25, Cisco set guidance of $1.16 to $1.18 in non‑GAAP net income per share on revenue of $16.7 billion to $16.9 billion, comparing to analyst expectations cited by Seeking Alpha of $1.07 per share on $15.82 billion in revenue.

Workforce reduction to reallocate resources toward silicon, optics, security and AI

Cisco announced plans to lay off 4,000 employees — nearly 5% of its workforce — as the firm shifts investment toward higher‑demand areas. Cisco employed 86,200 people globally as of July 26, 2025, according to regulatory filings. CFO Mark Patterson told investors the layoffs "were about dedicating more resources to silicon, optics, security and AI rather than trying to save money," and the company expects to spend up to $1 billion on severance and other one‑time termination benefits through mid‑2027.

The announcement comes in a wave of tech job reductions tied to refocusing around AI: the source notes that Cloudflare cut more than 1,100 workers (about 20% of its staff) and Arctic Wolf laid off 250 employees (about 7% of its workforce) in the same period as companies reallocate resources.

Acquisitions target agentic governance and mult i‑agent observability

Despite the layoffs, Cisco completed two acquisitions during the quarter: Astrix, described as a non‑human identity security firm, and Galileo, an AI observability company. Cisco said it hopes Astrix will supply capabilities for governance of AI agents and for threat detection and response tied to agentic technology. Galileo’s technology is intended to provide real‑time observability and guardrails for multi‑agent systems, according to the source.

Those moves are paired with Cisco’s participation in Anthropic’s Project Glasswing and OpenAI’s Trusted Access for Cyber program, positioning the company "at the center" of how the source describes AI being integrated into enterprise security and infrastructure.

What this means for technologists, procurement leaders, and investors

  • Technologists and security teams: Expect heightened attention to equipment that has reached end of support. Cisco’s framing makes testing, accelerated patching and replacement of unpatchable devices central to how its AI pilots are being justified.
  • Procurement leaders and enterprise customers: The company’s messaging and product bets — plus its work with Anthropic and OpenAI programs — are meant to nudge upgrades and new purchases; Robbins said Mythos can incentivize customers to move off unsupported infrastructure even if immediate orders tied to Mythos were not "meaningful" in the most recent quarter.
  • Investors and market participants: Strong AI‑infrastructure order receipts and raised guidance correlate with the stock’s after‑hours surge; Cisco’s guidance for the coming fiscal quarter outpaces the analyst numbers cited in the source and frames the company’s near‑term financial expectations around AI demand.

Cisco has paired a bullish AI‑infrastructure narrative with a painful workforce reallocation and targeted acquisitions. The core tension Robbins laid out is straightforward: AI tools can speed testing and patching, but they also expose the limits of legacy hardware and software that "can’t be patched" — a problem the company is now pitching as both a security risk and a revenue driver. Whether that calculus produces sustained hardware replacement cycles, new security contracts, or further operational upheaval for Cisco and its customers remains the next concrete test.

Original story