Most device refresh plans are built a year in advance, while DRAM prices move on a much shorter cycle. Buying a year’s worth of devices now, at a price you know today, lowers your exposure to whatever the market does next.
Here's what's driving these price changes and what to work out before you commit.
Why do DRAM prices keep changing?
A refresh plan delivers devices in batches across the year, and each batch pays whatever DRAM costs the month it ships. AI data center demand continues to absorb memory supply, and new fabrication capacity takes years to come online, so supply can't respond on a quarterly cadence. That mismatch means a plan built on last year's assumptions can land well above the budget it was built against.
Three questions to settle before you buy ahead
Buying ahead is a bigger commitment than a normal refresh, and three questions decide whether it holds up.
Where will the devices go? A year of hardware needs somewhere to sit between purchase and deployment, and some organizations don't have the necessary storage space or conditions. How much space you need depends on the delivery schedule: devices deployed in month two need storage for weeks, devices deployed in month ten need it for most of the year.
Which configurations should you order? One standard build rarely fits every team. What people need depends on the work they do, and at volume, that choice becomes the machine most of your workforce uses every day. Usage data shows where the differences are, including which users have outgrown what they're working on.
How do you make the case for the timing? Moving spend earlier means justifying the volume without the calendar as your evidence. Age alone won't carry it, since a three-year-old machine in light use is in better shape than one that's been pushed hard since day one. Device data shows which machines should be retired, and which can move to a role with lower demands.
What your device data can answer
Most organizations are already collecting the information needed to answer all three questions. That information is telemetry: the performance, usage, and condition data devices report while people work on them. Read across the device estate, it turns each of those three questions into a number.
Telemetry shows how hard a machine has been worked, not only how long it has been deployed. That separates devices that should be retired from those with life left, giving you a replacement volume built from the fleet instead of the calendar.
Usage patterns by role show which configurations people are working in, where the standard build falls short, and which users need more than their current machine gives them.
With volume and condition settled, the refresh can be sequenced: which devices go first, when each batch arrives, and how long each one needs to be stored before deployment.
Where a partner fits
Turning telemetry into a buying plan is ongoing work, and the hardware has to sit somewhere in the meantime. Most technology teams have enough on their plate without adding a warehouse operation. Unisys Device Subscription Service handles device life cycle management end to end: telemetry analytics that match devices to roles, warehousing that holds each batch until it deploys, and redeployment for machines that still have life in them. Smart purchasing works with any procurement method.
Want to explore how to build your refresh plan based on evidence? Start with a rapid value assessment for an intelligent device refresh.
