Supplier roadmaps point one way: higher density, newer standards, faster interfaces. Over the past twelve months, more engineers have moved the other way.

Key findings

• According to EETech's 2026 Memory Market Study, conducted with MEMPHIS Electronic, 27% of engineers shifted to older or more readily available memory technologies in the past 12 months, while 13% moved to higher-density devices.
• When engineers described next year's biggest memory challenge in their own words, those who named a specific device almost all named a legacy or mature part, such as DDR3, DDR4, LPDDR4 or low-density eMMC.
• Only 7% of engineers standardized on fewer memory families, the least chosen action in the study. 39% qualified alternative memory devices or suppliers.

Are engineers moving to older memory technologies?

More engineers moved to older memory than to newer. In EETech's 2026 Memory Market Study, in partnership with MEMPHIS Electronic, 27% of engineers shifted to older or more readily available memory technologies over the past 12 months, while 13% moved to higher-density devices. More than twice as many chose older or more available parts as chose higher density, one of the two clearest differences in the study.

Which memory parts do engineers name as their biggest challenges?

Among engineers who named a specific device, nearly all named legacy or mature parts. Asked to describe the biggest memory challenge they expect next year, most engineers raised availability (60%) or price (39%) in general terms, and 11% raised shortages of specific legacy parts. The devices named were: DDR3, DDR4, LPDDR4, low-density eMMC, 3.3V OctoSPI flash, and PSRAM. One named DDR5. None named HBM. This reflects the engineers in this study, a base weighted toward industrial and long-lifecycle markets, rather than a measurement of the whole memory market.

"DDR4 supply has completely collapsed. We can't even get quotes anymore, just watching our inventory drain day by day."

Where do long product lifecycles fit in?

They shape the context. More than half of respondents (54%) design for markets built on long product lifecycles and demanding qualification, led by industrial automation and manufacturing (28%), alongside aerospace and defense, automotive, transportation, utilities, and medical devices. These products are expected to stay in production longer than the components available to build them.

"Our products have a ten-year lifecycle minimum, but memory fabs want to EOL parts every two years. The timelines just don't match."

The pressure is not limited to those markets. The study found no measurable difference between long-lifecycle markets and the rest of the sample in how far engineers have had to change their designs.

How are engineers responding to memory supply risk?

They are choosing optionality over standardization. Standardizing on fewer memory families, the conventional route to simpler qualification and lower cost, was the least chosen action in the study, at 7%. By contrast, 39% qualified alternative memory devices or suppliers, 31% built more design flexibility to accommodate different memory options, and 31% reduced their reliance on a single supplier.

Each of those choices is an opening for a supplier that is ready to be qualified. Lowering the engineering cost of qualifying your parts is covered in the third post in this series.

Why do design-ins made during the memory shortage matter?

Because they outlast the shortage. 35% of engineers redesigned products around the memory components they could obtain, and 11% report their organization discontinued a product outright. Redesign is slow, expensive, and rarely a first resort, and the part chosen governs demand for that product's production life.

That cuts both ways. A supplier who is present now earns a socket that is unlikely to be revisited when the market settles. A supplier who is absent has not deferred a sale. It risks being designed out for years.

What should memory suppliers and distributors do about older and mature parts?

Suppliers: explicit longevity programs and generous last-time-buy windows on the devices roadmaps are leaving behind. Engineers designing long-lifecycle products need a commitment they can design around.

Distributors: stocking depth and inventory visibility on exactly those parts. With availability named as a top sourcing concern by 69% of engineers, as covered in why engineers rank availability above performance, the distributor who can show stock on the older and mature parts engineers are struggling to source has the stronger conversation.

Both: make legacy and mature coverage easy to find. Engineers are qualifying alternatives and moving to new sources, so coverage needs to be visible where they search.

Frequently asked questions

Which memory parts do engineers name as their biggest challenges in 2026?

Among engineers who named a specific device in open-ended responses to EETech Research's 2026 Memory Market Study, nearly all named DDR3, DDR4, LPDDR4, low-density eMMC, 3.3V OctoSPI flash or PSRAM. One named DDR5 and none named HBM.

How many engineers moved to older memory technologies?

27% of engineers shifted to older or more readily available memory in the past 12 months, compared with 13% who moved to higher-density devices.

Are engineers standardizing on fewer memory families?

Rarely. Only 7% standardized on fewer memory families. 39% qualified alternative devices or suppliers and 31% increased design flexibility.

How many engineers redesigned products because of the memory market?

35% redesigned products around the memory components they could obtain, and 11% report their organization discontinued a product.

The full 2026 Memory Market Study covers how engineers have changed their designs, what they expect from the next twelve months, and recommendations for memory suppliers and distributors.

Download the 2026 Memory Market Study

‍