A PC refresh cycle is the schedule on which an organization replaces its computers, and for PCs and laptops it commonly runs three to four years, according to the European Commission’s Joint Research Centre. A cycle that replaces every machine on that schedule spends budget on devices that still work, and it keeps some that should have gone first. A better cycle uses age only to decide which devices to review, and lets evidence decide which ones to replace: whether the hardware meets the requirements the organization runs on, whether the device has stayed stable, and whether IT can fix its problems without new hardware.
Why the Calendar Makes a Poor PC Refresh Cycle
The cost of getting this wrong is rising. IDC forecasts average PC selling prices to grow 17% in 2026, driven by a memory shortage it does not expect to ease before the end of 2027. In the same forecast, IDC expects global PC shipments to fall 11.3% this year. Every unnecessary replacement now costs more than it did a year ago.
Age is a proxy, and a weak one. Two laptops bought on the same day can end up in very different condition: one runs as well as it did on day one, the other crashes twice a week and sits in the helpdesk queue. A calendar-based cycle treats them the same. It replaces the healthy one early and leaves the failing one in service until its date comes up.
The mistake costs in both directions. Replace too early, and the budget buys hardware the organization did not need, and the emissions come with it: the JRC report finds that manufacturing is the dominant source of a computer’s greenhouse gas impact, and cites research estimating that extending laptop lifetime from three to five years can cut an organization’s related emissions by 37%. Replace too late, and IT pays in tickets, reinstalls, and users who work around a machine they no longer trust.
Slow Is Not the Same as Worn Out
Many complaints about an old PC are really complaints about a slow or unstable one. The causes differ, and only some of them call for new hardware.
IT can fix the first kind. A faulty driver, a failed update, or a single program that keeps crashing can make a capable machine feel worn out. Windows records these problems in its event logs, so the evidence usually exists before anyone opens a ticket. Fix the cause, and the device can serve for years more.
The second kind sits in the hardware itself. Windows 11 requires a TPM version 2.0 and a processor on Microsoft’s list of supported processors. The processor is a fixed limit: no setting, script, or reinstall changes its generation. The TPM is worth a second look before anyone writes the device off, because Microsoft notes that many PCs capable of TPM 2.0 ship with it turned off in the firmware settings, and some vendors provide a firmware update for it. Windows 11 is the clearest example right now, but the same logic holds for any requirement the hardware cannot grow into.
Some wear sits in between. A battery that no longer lasts a working day makes a laptop feel finished, but IT can replace the battery and keep the machine. The same often goes for a full or failing drive, or too little memory in a device with room for more.
Three Questions Before You Replace a PC
Does the Hardware Meet the Requirements?
Start with the components, not the purchase date. Check the processor model and the TPM version. On laptops, also check how much of the battery’s design capacity remains: that decides between a repair and a replacement.
Two Windows 10 devices in one of CapaSystems’ own CapaOne environments show why the details matter. One is a Lenovo ThinkPad X240 with a 4th-generation Intel Core i5-4210U and a TPM 1.2, and it looks as old as it is. The other is an HP EliteDesk 800 G3 Mini with a TPM 2.0, and on a TPM check alone it looks ready for Windows 11. It is not: its 7th-generation Intel Core i7-7700 is missing from Microsoft’s supported list, just like the ThinkPad’s processor. Both get the same verdict, replace rather than upgrade, and only the processor model reveals it for the EliteDesk.
Has the Device Stayed Stable?
Look at stability over weeks, not at the most recent complaint. A device whose critical events and errors rose after a specific change has a cause worth finding. A device that has logged unexpected shutdowns for months, with no software change behind them, points toward the hardware. Event logs alone do not prove a hardware fault, so run the vendor’s hardware diagnostics for storage, memory, and temperature before deciding.
Can IT Fix the Cause Without New Hardware?
If the instability traces back to software, fix it and keep the device. Our guide to endpoint performance monitoring walks through how to trace a recurring Windows error to its source across the fleet. If a single worn part explains the problem, repair it. If nothing short of new hardware explains it, or the processor misses a requirement, the device belongs in the replacement plan.
Building a PC Refresh Cycle on Device Data
A PC refresh cycle built on evidence still uses age, but only to decide which devices to review. Set the review age, list the devices that reach it, and give each one a verdict:
| What the review finds | Verdict |
|---|---|
| Hardware meets the requirements, and the device is stable | Keep |
| Hardware meets the requirements, but software causes errors | Fix the software |
| TPM 2.0 is available but turned off in the firmware | Enable it |
| A worn part, such as the battery, but the rest holds up | Repair |
| The processor misses a requirement the organization depends on | Replace |
| Still unstable after software fixes, and diagnostics find a fault no single part explains | Replace |
Devices that keep running come up for review again next year. Among the devices to replace, the ones that are both unsupported and unstable go first.
The same review gives IT a stronger case for the budget. Instead of a request for a fixed number of new PCs, leadership sees which devices IT plans to replace and why each one qualifies: a processor outside the supported list, or months of instability that neither a software fix nor a repair solved. Every device that passes the review defers a purchase by at least a year and spreads the cost of the fleet over more years of use.
How CapaOne Shows the Evidence
The CapaOne Endpoint Management Platform brings the evidence for each Windows device into one place. Open any Windows endpoint, and its hardware inventory shows the processor model, the TPM version, and the battery’s current full-charge capacity against its design capacity. Experience Monitor collects the critical events, errors, and warnings from the Windows system and application logs, gives every endpoint a reliability score from 1 to 10, and keeps up to three months of score history, so IT can see whether a device has stayed stable or whether a fix held.
The useful question about the next round of PC purchases is not how old the machines are. It is which machines can still carry the work, and the devices already hold that answer.
Book a demo to see the hardware inventory and reliability scores for your own Windows endpoints.
