26 February 2026

Chris Noon. Director of Solution Engineering, International for Alkira

For IT leaders, the math no longer adds up. While budgets remain under intense scrutiny, the cost of the "boxes" required to run a modern business is climbing at an alarming rate. We are witnessing a structural shift in the economics of networking—one that requires a fundamental rethink of how we build and consume infrastructure.

Price Hikes

The current inflationary trend isn't limited to a single vendor or region. Data from throughout 2025 and into early 2026 reveals a coordinated upward pressure across the entire hardware stack:

• Core Routing & Switching: Industry reporting indicates Cisco made broad pricing adjustments in 2025, commonly cited at ~3.4% on hardware effective September 13, 2025, with similar uplifts for technical services in early October 2025. Separately, some industry coverage and tariff-impact analyses suggest that select high-end platforms could see larger, quote-dependent increases in the ~10–15% range under certain tariff exposure scenarios, which should be treated as scenario/estimate-based, not universally applied list pricing.
• Wireless Infrastructure: Industry reporting suggests WLAN pricing can be sensitive to tariff and component-cost dynamics, with high single-digit increases showing up in some contexts. Avoid framing this as a uniform, vendor-announced 8–12% increase across the board unless you can cite a specific, attributable vendor notice.
• The “AI tax” on memory (servers): The clearest pressure point is server memory. Industry research and reporting indicate server-grade DDR5 RDIMM pricing could approach ~2× by end of 2026 versus early 2025 amid AI-driven demand. Some reporting also notes meaningful server quote impacts in certain configurations, including reported increases up to ~25% in some cases.

Why the Old Playbook is Failing

For decades, the standard operating procedure was simple: forecast capacity, secure capital expenditure (CapEx), and "rack and stack." This model relied on three assumptions that are no longer true: stable prices, predictable lead times, and available talent.

Today, tariffs on Chinese and Mexican manufacturing have turned procurement into a geopolitical gamble. Lead times remain elastic, and the skilled engineers required to configure and patch this increasingly expensive hardware are harder to find. When you factor in the "hidden" costs—power, cooling, rack space, and the opportunity cost of having your best people managing firmware updates instead of architecting business solutions—the total cost of ownership (TCO) is reaching a breaking point.

Best Practices for the New Reality

As we navigate this "cost crisis," networking professionals must shift from being hardware gatekeepers to service orchestrators. Here is how leading organizations are adapting:

1. Audit for "Ghost" Infrastructure: Before approving a new hardware spend, perform a rigorous audit of current utilization. Industry data suggests that 15-20% of enterprise servers and networking ports sit idle, yet still consume power and maintenance budget. Consolidating workloads can defer the need for a costly refresh.

2. Bridge the CapEx-OpEx Divide: The volatility of hardware pricing makes long-term depreciation schedules a liability. Shifting toward an Operational Expenditure (OpEx) model allows for financial agility. By consuming infrastructure as a service, organizations can align their spending with actual usage, effectively "outsourcing" the risk of tariff-driven price spikes to the service provider.

3. Prioritize "Zero-Touch" and Automation: If you must buy hardware, prioritize vendors that offer robust automation. The true cost of a switch isn't the PO price; it's the 40 man-hours required to deploy and secure it over its lifespan. Look for platforms that treat hardware as a commodity and management as a centralized, software-defined function.

4. Decouple Growth from Hardware: The most successful IT leaders in 2026 are those who have decoupled their ability to scale from their ability to procure physical boxes. Whether it’s expanding into a new global region or spinning up a cloud-bursting environment, the "physical-first" approach is too slow and too expensive.

The Path Forward

At some point, every IT leader has to ask: Why are we still in the business of owning power plants? We don't build our own electrical grids, and we've largely moved away from building our own data centers. Networking is the final frontier of this evolution.

When the network is delivered as a global, on-demand service, the complexities of chip shortages, shipping logistics, and 25% tariff hikes become the provider's problem to solve, not yours.

By moving to a cloud-native networking model, you eliminate the need for massive upfront capital outlays. You gain instant global reach without waiting for a shipping container to clear customs. Most importantly, you free your engineering team to focus on the applications and security policies that actually drive revenue.

The organizations that thrive in the coming years won't be the ones with the largest hardware budgets. They will be the ones that had the foresight to stop buying boxes and start consuming the network as the utility it was always meant to be.