Blog The Ram Shortage Is Real.
The Panic Doesn’t Have To Be.

How to navigate the memory crunch without overspending — or falling behind.

The headlines are hard to miss. RAM stick prices are climbing. OEMs are warning customers about potential price hikes, with some experts predicting a 30% increase in device prices this year. Retailers are comparing DRAM chips to market-price lobster. And analysts are projecting the squeeze could last through Q4 2027 — possibly into 2028.

If you’re an IT or finance leader, the question landing on your desk is the same one every CFO is weighing: Do we buy now, or wait this out?

Our answer, after ongoing conversations with IT leaders navigating this exact decision: Stop framing it as a binary. The smartest organizations aren’t panic-buying, and they aren’t waiting either. They’re doing something more disciplined in the middle.

What’s actually happening?

The shortage is genuine. But it’s less a supply collapse and more a reallocation of priorities. RAM manufacturers are making a rational business decision: AI data centers represent a booming, high-margin market, and finite manufacturing capacity is being redirected to meet this demand. PC RAM — a commoditized, thin-margin product in a race-to-the-bottom pricing environment — is getting deprioritized. Whatever PC RAM remains is seeing hiked prices as a result.

There’s also a hype cycle layered on top. Some of the price action reflects opportunism — companies making hay while the sun shines — and there’s a fair question about when the AI bubble bursts. But the underlying scarcity is real. Expect elevated PC pricing to persist for at least the next 12 to 24 months, and potentially through 2028.

So, is this the new normal? For the foreseeable future, yes.

The old refresh playbook is obsolete.

Here’s the reframe most organizations are missing: the memory shortage isn’t just a procurement problem. It’s an invitation to retire a lifecycle model that stopped making sense several years ago.

The traditional three-year PC refresh cycle was built for an era when devices had multiple moving parts, mechanical failure points, and workloads that routinely outgrew their hardware. That era is over. Modern laptops have essentially one moving part — a fan. An i5 or i7 with 16GB of RAM has enough surplus capacity to handle the workload of the average employee — email, browsing, slides, tables — for years longer than most refresh policies assume.

Four to five years is the new acceptable refresh window for the average user.

Six or seven is too long — things genuinely start falling off. But three is leaving money on the table. IT developers, designers, and other power users still warrant a faster cycle. Everyone else doesn’t.

From calendar-based to needs-based refresh

The single highest-leverage shift a business can make right now is moving away from, “This laptop is four years old, time for a new one,” and toward, “What does the data say about this device?”

That means instrumenting your fleet. Between your MDM platform — Intune, Jamf, SCCM — and a Digital Employee Experience (DEX) agent deployed to the endpoint, you can gather deeper analytics on Mac and Windows devices: average CPU and RAM consumption, GPU usage, battery cycles, and overall health and performance.

With that telemetry, three categories emerge clearly: Retain, cascade, and refresh.

A light user on healthy four-year-old hardware? Retain. A power user whose machine is choking on their workload? Refresh — then, cascade their old device to someone with lighter needs. Reimage a Windows PC that’s been bogged down with constant patches, and you can often get another 6 to 12 months out of it.

This is how you protect the budget without protecting the wrong assets.

Persona mapping: The discipline most teams skip

“Everyone gets the same laptop” is an expensive policy in a normal year. In a shortage year, it’s untenable.

The sweet spot most organizations land on is 10 to 15 distinct personas — enough granularity to right-size hardware against actual workload, depending on whether the business leans toward manufacturing, data engineering, or something else. Most companies say they want to do persona-based provisioning. Few are doing it well.

The blocker is almost always the same: They don’t have good data on who their users are, which devices they have, and how those devices get used. Fix that data problem, and the procurement problem gets smaller.

If you’re buying in the next 12 months, buy now.

For the devices you’ve already identified as genuine refresh candidates, the calculus is clear. Every OEM is signaling prices will climb as the year progresses. If you know what you need to deploy in the next 6 to 12 months, lock it in at today’s prices.

Partner with a provider who can warehouse the inventory until deployment. The certainty is worth more than the marginal hope of a price drop that analysts aren’t forecasting.

One nuance worth flagging: Not long ago, the smart future-proofing move was specifying 32GB of RAM as standard to extend device life to five or six years. That math has shifted. With 32GB machines now significantly more expensive, part of that strategy is worth revisiting.

What about security?

A fair objection: Older devices are security risks, right?

Mostly, no. As long as a device can run current releases of the operating system with the latest security patches applied, age alone isn’t the vulnerability.

The weak spot in endpoint security is the user, not the device. Keep users on VPNs. Enforce strong passwords and multifactor authentication. Retire hardware that can no longer run supported, patched OS versions. Beyond that, a well-maintained older laptop is not meaningfully more vulnerable than a new one.

A note on Apple

Apple hasn’t announced RAM-related price increases. Don’t mistake that for immunity. With a new model lineup on the way, any cost pressure will likely be absorbed into launch pricing rather than disclosed as a separate adjustment — a smart move that lets them avoid spooking the market.

If Apple is part of your mix, plan for higher effective costs on the next refresh regardless of what the launch messaging says.

The leader’s playbook

If you’re walking into a budget conversation this quarter, here’s the argument:

1. Reject the binary. “Buy everything now” and “wait it out” are both extremes. The answer is targeted action informed by data.
2. Extend the refresh cycle to 4–5 years for standard users. Reserve faster cycles for developers, designers, and other power users.
3. Instrument the fleet. Use MDM and DEX telemetry to separate retain, cascade, and refresh candidates.
4. Lock in near-term needs at today’s prices. For the 12-month refresh queue, buy now and warehouse.
5. Right-size with personas. Stop buying the same spec for everyone.
6. Cascade internally. A reimage can unlock 6 to 12 months of additional life for lighter users.
7. Revisit your 32GB default. What made sense as future-proofing last year may not pencil out today.
8. Don’t wait for devices to fail. Needs-based refresh means anticipation and predictive failure data — not running hardware until it won’t turn on.

The bottom line

This isn’t quite the crisis some fear it is. It’s a constraint — and constraints tend to reward the disciplined.

Organizations that come out of the next 24 months in the best shape won’t be the ones that bought the most devices or the fewest. They’ll be the ones who retired calendar-based refresh, built a real data layer under their fleet decisions, and treated this moment as the forcing function it is.

The RAM shortage will ease, eventually. The discipline you build around device strategy now will pay off long after it does.

Insight’s Flex for Devices program combines fleet telemetry, persona mapping, and procurement flexibility to help organizations take a needs-based approach to device lifecycle — refreshing what needs refreshing, cascading what can cascade, and making smarter capital decisions in a market that’s anything but predictable.