Rising memory prices: what UK organisations need to know in 2026

Earlier this year, we explored the wider pricing pressures affecting IT hardware and how UK businesses could plan ahead. Since then, memory has become one of the clearest drivers of rising device and infrastructure costs.

Organisations receiving new hardware quotes have been experiencing ever-increasing prices, shorter quote-validity periods and a wider cost difference between standard and higher-memory configurations.

Understanding what is happening can help IT and procurement teams make informed decisions before limited supply and changing prices begin to dictate their options.

Why memory prices are rising

DRAM provides the working memory used by laptops, desktops and servers, while NAND Flash is used in solid-state storage.

Both are essential components across workplace devices, servers and storage systems.

TrendForce forecast conventional DRAM contract prices to rise by 58–63% during the second quarter of 2026. Although the rate of increase is expected to slow, prices are still forecast to rise by a further 13–18% during the third quarter, alongside a 10–15% increase in NAND Flash prices.^[1][2]

Gartner estimates that combined DRAM and SSD prices could be around 130% higher by the end of 2026 than they were in 2025. It expects these component pressures to contribute to a 17% increase in PC prices.^[3]

Using a separate market forecast, IDC projects that average PC selling prices will rise by 18.3% during 2026, with no meaningful relief from the memory shortage expected before the end of 2027.^[4]

Component-price increases do not transfer directly to the final price of every device. Manufacturers can respond through price changes, revised configurations, reduced discounting or a combination of these measures.

This means an organisation may pay more for the same specification, or find that a device offered at last year’s price now includes less memory or storage, a possible form of technology shrinkflation.

AI infrastructure is changing how capacity is allocated

The current pressure is closely connected to the continued growth of AI infrastructure.

AI servers require large quantities of high-bandwidth memory, or HBM. Memory manufacturers are increasing production of these higher-value products to meet demand from data centres and cloud providers.

That shift places additional pressure on the capacity available for conventional memory used in business laptops, desktops and other systems. TrendForce reports that ongoing capacity allocation towards server applications is reducing the supply available for PC DRAM.^[2]

This is significant because new semiconductor manufacturing capacity takes years to plan, build and bring into production.

Micron reported in June that industry demand for DRAM and NAND continued to significantly exceed supply. The company expects tight conditions to persist beyond 2027 because of AI-driven demand and structural supply constraints.^[5]

There may eventually be gradual improvements in supply, although manufacturers currently have limited visibility over when the market will fully rebalance.

Memory now represents a much larger share of device costs

The effect is already visible within hardware manufacturers’ own costs.

During its February 2026 earnings call, HP reported that memory and storage had previously represented approximately 15–18% of its PC bill of materials. For its 2026 financial year, the company estimated that figure would rise to around 35%.^[6]

When such a large proportion of a device’s cost is concentrated in memory and storage, manufacturers have fewer ways to absorb the increase.

This can affect:

• Laptop and desktop prices
• The cost of moving from 16GB to 32GB of memory
• Larger SSD configurations
• Server and storage projects
• The availability of particular models or specifications
• The length of time quoted prices can be held

Organisations planning purchases using figures obtained earlier in the year may therefore find that the same budget no longer covers the same quantity or specification.

The timing creates an additional challenge

Many UK organisations are still completing their transition from Windows 10.

Standard support for Windows 10 ended on 14 October 2025. Microsoft offers a paid Extended Security Updates programme for organisations that need additional time to migrate, although it is intended to provide a temporary bridge rather than a long-term replacement for moving to a supported operating system.^[7][8]

This means a significant number of device-refresh programmes are now taking place during a period of higher memory costs and constrained supply.

New AI-capable devices may also carry higher minimum requirements. Microsoft’s specification for Copilot+ PCs includes at least 16GB of DDR5 or LPDDR5 memory and 256GB of storage.^[9]

For organisations preparing for greater use of local AI tools, buying devices with insufficient memory could limit their useful lifespan. Applying a higher specification to every user can also increase fleet costs unnecessarily.

The specification decision needs to be based on genuine user requirements.

How organisations can respond

Review planned purchases earlier

Map expected device, server and storage requirements over the next 12 to 18 months.

This provides time to compare options, assess availability and decide which purchases should be brought forward. Waiting until equipment becomes urgent usually reduces the choice of suitable configurations and delivery schedules.

Treat quote validity as part of the decision

Ask how long pricing can be held and make sure internal approval processes reflect that timescale.

A quote that remains in an approval queue for several weeks may need to be repriced before the order can be placed.

Match specifications to user requirements

Review which employees and workloads genuinely require higher-memory configurations.

Some users may need 32GB or more for technical applications, data analysis, creative software or local AI workloads. Others may be well served by a carefully selected 16GB device.

This helps protect performance without applying unnecessary costs across an entire fleet.

Consider selective life extension

Some existing devices may continue to meet operational and security requirements with targeted upgrades or maintenance.

Extending suitable assets can help spread purchasing across more than one budget period. Each device should still be assessed for Windows 11 compatibility, security, performance and ongoing support.

Plan procurement and deployment together

A phased programme can provide more control than one large purchase made at the prevailing market price.

Organisations may be able to agree pricing, secure suitable stock and schedule deployment in stages around budget cycles, operational priorities and available resources.

Planning your next refresh with VitrX

VitrX supports organisations with technology procurement, workplace devices, infrastructure, logistics and deployment.

We can help you review your current estate, identify upcoming requirements and match device specifications to different users and workloads. We can also explore available vendor options, current lead times and phased procurement approaches.

Nobody can predict exactly when memory pricing will settle. Current manufacturer and analyst guidance points towards continued pressure through 2027, making early visibility increasingly valuable.^[4][5]

If your next device refresh, server upgrade or education technology programme is planned within the next 18 months, now is a sensible time to review the assumptions behind it.

Planning a device refresh or infrastructure project? Speak to VitrX about current pricing, availability and the options for your estate.