The memory demand crunch: creating a device strategy that meets the challenge

Datacom

Enterprise IT leaders building 2026 budgets face a hardware procurement challenge unlike anything since the pandemic-era chip shortage. But unlike 2020–2021, this is not a short-term event. The AI boom has triggered an ongoing structural shift in memory supply.

AI systems consume memory at 10 to 20 times the rate of traditional PCs, and the hyperscalers building infrastructure at scale will always be first in line. That leaves enterprises competing for an increasingly constrained supply of the memory that powers their everyday business devices, and prices are likely to rise with each passing quarter.

There are very few RAM manufacturers globally, and their production lines can produce either DRAM, which is the memory used in devices such as PCs, laptops, servers, and smartphones, or High Bandwidth Memory (HBM), which powers AI accelerators. Retooling between the two different products is capital-intensive and time-consuming.

The surge in AI adoption has triggered an unprecedented spike in orders for HBM, prompting manufacturers to shift their production focus toward this newer, more profitable technology. As a result, DRAM supply has contracted, and prices have risen sharply in proportion to the amount of memory a given product requires. 

The cost of waiting

The market is already sending clear signals that patience is not a strategy. Quote validity windows are shortening, and suppliers are inserting repricing provisions into contracts, which means the price agreed upon today may not hold if manufacturers’ costs shift before delivery. Enterprise IT is already seeing substantial price increases. A business notebook that cost around $2,000 twelve months ago may now approach $3,000. Further increases are projected throughout the rest of the year. Budget assumptions based on today’s hardware costs will quickly become outdated. 

“Today is the best price you’re likely to get for a long time,” said Chris Diaz-Poff, Team Lead for EUC and Infrastructure Solutions at Datacom. “If you have a forecasted need for devices over the next 12 to 24 months, consider early procurement.”

To manage costs, CIOs can take advantage of manufacturer-provided financing and leasing arrangements, options that can include hardware buyback programs, payment deferrals, and extended warranties. In this way, organisations can lock in today’s pricing while preserving budget flexibility.

Datacom maintains close engagement with global vendors and distributors to monitor the daily shifts in availability, pricing, and allocation that define the current market. Understanding these fluctuations enables a more strategic approach to refresh planning, allowing potential supply risks to be identified and addressed well in advance. When preferred configurations become constrained, having a broader view of the vendor landscape makes it easier to identify viable alternatives without compromising performance. Ultimately, early engagement and forward-looking procurement roadmaps are the most effective tools for securing price stability and preventing supply chain volatility from causing operational disruption.

Platform strategy also matters. For example, transitioning to Windows Autopilot deployment and Microsoft Intune device management gives IT hardware-agnostic provisioning capability with zero-touch deployment that works regardless of the specific device being provisioned. This enables IT teams to source the best-value hardware available without adding deployment overhead, all while reducing exposure to single-vendor supply chain risk.

A smarter foundation: AMD Ryzen AI and lifecycle management

For organisations looking to stretch the value of every memory gigabyte they purchase, the choice of processor architecture matters. AMD Ryzen AI processors consolidate CPU and GPU memory requirements into a single chip, a design that becomes particularly powerful in the Strix Halo lineup. Unlike conventional platforms, Strix Halo devices can dynamically allocate memory between system and graphics functions, enabling workstation-class visual performance without the dedicated GPU memory that would otherwise significantly increase costs. For organisations procuring AI-capable workstations, this translates to material savings per unit.

Datacom works with organisations to understand both current market conditions and forward-looking cost trajectories, enabling data-driven, cost-rational procurement decisions, rather than reactive ones. In a market where the structural pressures on memory supply are unlikely to ease anytime soon, that kind of informed, strategic approach to the device lifecycle is no longer a nice-to-have. 

For CIOs preparing their 2026 budgets, it may be the most important investment decision they make this year.

To learn more about how Datacom and AMD Ryzen can help CIOs overcome the ongoing memory challenge, click here.