Cisco CEO Chuck Robbins says customers are rushing to replace unsupported networking equipment after Anthropic's Mythos bug-finding model made older devices too risky to keep running. Speaking on the company's Q4 earnings call, Robbins said buyers are moving fast and reallocating security budgets to pay for replacements, worried that AI-powered vulnerability scanners will expose flaws in gear that no longer receives patches. He calls it "The Mythos Effect" and believes it's fueling a broader "supercycle" of network spending alongside preparation for quantum computing and the rollout of agentic AI.
Cisco posted revenue of $17.3 billion for Q4 and $63.3 billion for the full year, climbing 17 percent and 12 percent respectively. Net income jumped 51 percent to $3.9 billion in the fourth quarter, while the annual figure of $13.27 billion represented a 30 percent increase. Robbins noted the company achieved its highest revenue, operating margin, and earnings per employee in three decades. CFO Mark Patterson said the company isn't facing significant lead time delays for customers ordering replacement equipment, unlike competitors who've reported supply constraints.
According to Robbins, one customer—a CEO running a major U.S. manufacturer—reached out early "in the Mythos wave" saying their team needed to swap out gear past its last day of support. "We've seen the pipeline increase meaningfully as a result of Mythos, which is really showing up as a network refresh," Robbins said. The CEO also revealed that Cisco resolved 145,000 support cases "entirely by AI with zero human intervention" during fiscal year 2026, and that the company's internal AI assistant called Circuit handled over 75 million prompts in Q4 alone.
The Mythos Effect works because Anthropic's model has proven so capable at uncovering security flaws that vendors and open source projects are now discovering more vulnerabilities and issuing more patches than before. That success has made customers fearful that unsupported devices—ones no longer receiving those patches—will become dangerously exposed once models like Mythos identify exploitable weaknesses. Robbins thinks some buyers are shifting money from security budgets rather than networking allocations to pay for the upgrades, a move he's happy to accommodate regardless of the funding source. He expects the spending wave to continue through fiscal 2027, with all key financial metrics improving further as customers refresh infrastructure to handle AI workloads and protect against quantum-era threats.
Robbins told investors he expects revenue, operating margin, and earnings per employee to keep climbing in the coming fiscal year as the network refresh cycle continues. Circuit, Cisco's in-house AI assistant, runs on what the company calls its Secure AI Factory infrastructure, which improves GPU efficiency and routes tasks to the right large language model to control token usage. Despite the strong results, Cisco's share price spiked immediately after the earnings announcement before settling four percent below its previous close, suggesting investors remain uncertain about the durability of the AI-driven demand surge. The shift from IT discretionary spending to security-driven urgency may redefine how enterprises budget for infrastructure in an era where obsolescence carries not just operational risk but existential vulnerability. Companies that delay refreshes could find themselves caught between ballooning security insurance premiums and the cost of emergency replacements when legacy gear becomes a verified liability.