Cisco Hardware Price Increase: Memory Costs Surge [2026]

Cisco CFO Mark Patterson noted during the call that he hasn’t spoken to any customers willing to delay strategic investments. CEO Chuck Robbins echoed this sentiment, stating, "Customers get it. And while they may not like it, they understand that it’s a dynamic that we’re all dealing with."

It helps that Cisco is seeing its own AI boom. The company secured $2.1 billion in AI infrastructure orders from hyperscalers in the quarter—an $800 million jump from the previous quarter. This suggests that the very trend causing the memory shortage (AI adoption) is also fueling Cisco’s growth.

Is this just a Cisco problem?

Unfortunately for IT buyers, this isn’t isolated to one vendor. The memory crunch is an industry-wide event. Reports indicate that other major hardware players, including Dell, Lenovo, and HPE, are preparing similar price increases. Industry analysis suggests these competitors are facing pressure to raise prices on server hardware by up to 15% to cover the same DRAM crunch.

We are entering a period of hardware inflation where enterprise IT budgets will simply need to expand. The "capacity siphoning" toward AI chips means that standard infrastructure is becoming a scarcer resource. With Cisco’s networking revenue up 21% year-over-year, it is clear that a robust campus refresh cycle is underway, and companies are deciding that paying a premium is better than falling behind on their network modernization.

The Real Story

While Cisco frames this as a supply chain hiccup that customers "understand," the reality is that standard enterprise IT is now subsidizing the AI revolution. The memory industry has effectively decided that HBM for AI is the priority, relegating standard DRAM to second-tier status with scarcity pricing to match. For the next 12 to 18 months, CIOs should expect to pay an "AI Tax" on almost all infrastructure hardware—even if that hardware has nothing to do with artificial intelligence—simply because the raw materials are being diverted to feed the hyperscalers’ GPU clusters.

Source: Original Article