On November 6, 2026, market research firm TrendForce announced that it expects the use of reused memory in new smartphones to expand. The company estimates that reused memory will account for 3–4% of the smartphone market in 2026, rising to 7–9% in 2027. Adoption is currently concentrated mainly in 4G smartphones priced below 1,299 yuan that carry 4GB of LPDDR4X. Products that need older-generation components remain on the market, while new supply has become harder to secure, and the practice of reusing recovered memory chips is beginning to spread.
The reused memory in question is DRAM, which temporarily holds data, rather than the storage that keeps photos and the operating system. As Micron's explanation notes, NAND flash memory handles long-term data storage, while DRAM temporarily holds the data the CPU is processing. Why would a new smartphone need to be fitted with used DRAM? And how should its quality and reliability be judged? Behind the trend are the supply-and-demand structure of the memory market and the processing techniques used to make recovered chips reusable.
What Is "Reballing" for Reusing Used Chips?
"Reballing" is the process used to prepare reused memory. It means reattaching the small solder balls that electrically connect a semiconductor component to a circuit board. The connection method that places solder balls on the underside of a chip is called BGA (Ball Grid Array), and when a recovered component is mounted on another board, this joint has to be reworked. It is a process of rebuilding the connection points; it does not mean the memory chip itself is manufactured again in a semiconductor fab.
In the general reballing process published by Kei-All, the component is first removed from the board and the old solder remaining on it is cleared away. Appropriate solder balls are then placed on the component and heated, followed by cleaning and inspection. In other words, a recovered chip is not simply transferred to another board as it is; the connection points must be rebuilt. However, that document describes a general processing method and does not mean the company supplies the reused memory for the smartphones discussed here.
Nor does a reballed component necessarily mean a part that was used for a long time. According to an explanation from Pioneer Service Network, reballing is performed not only to reuse BGA parts but also to change from leaded to lead-free solder and for inspection and evaluation. The label "reballed" alone does not reveal how a component was recovered. For the reused memory in these smartphones, the specific breakdown of sources, such as used devices or discarded boards, has not been disclosed.
The key point is that reballing renews only the connection points; it does not change the memory chip's internal structure or its past usage history. For that reason, a description of the reballing process alone cannot tell us about differences in performance or lifespan compared with new parts. Nor is it appropriate to apply the lifespan constraints of NAND flash memory, which are based on the number of rewrites, directly to DRAM.
Why Budget Phones Can't Simply Move to Newer Memory Even as LPDDR4X Gets Pricier
Semiconductor components do not always get cheaper as they age. Even if some demand remains, a manufacturer's production cutbacks can push prices up as supply shrinks. In its July 17, 2025 research release, TrendForce reported that major memory makers in South Korea and the United States planned to scale back or end LPDDR4X production in 2025–2026. It also noted that many processors for low-priced smartphones support only LPDDR4X, so moving to a new memory generation is not easy.
These constraints can also be seen in SoC product specifications. For example, MediaTek's Helio G99, a SoC for 4G smartphones, supports LPDDR4X memory and the UFS 2.2 storage standard. Memory and storage are defined as separate specification items, so even if newer-generation DRAM is available on the market, it cannot necessarily be swapped in on an existing board. The Helio G99 is cited only as an example of supported standards and does not indicate that devices using this SoC adopt reused DRAM.
Moving from LPDDR4X to newer DRAM requires the SoC to support that standard, and in some cases board redesigns and fresh validation as well. Low-priced devices have difficulty passing the cost of such design changes on to the product price. Demand is therefore likely to remain for LPDDR4X, which can be procured while keeping existing designs. The combination of shrinking new supply and lingering demand is creating the economic incentive to reuse recovered DRAM.
The markets to which memory makers direct their production capacity are also changing. In its July–September results announced on October 28, 2026, SK hynix analyzed that supply is failing to keep pace with demand because of growing AI demand, the time needed to build manufacturing facilities, and the technical difficulty of advanced processes. The company said it would expand sales of DDR5 and LPDDR5 SOCAMM for AI servers and increase HBM4 production. This suggests major manufacturers are prioritizing the expansion of higher-value products, but AI demand alone cannot explain the shortage of mobile DRAM.
On the other hand, production of older-generation memory is not ending across the board. In its May 22, 2026 announcement about expanding capacity at its Manassas, Virginia plant, Micron said it would expand supply of DDR4 and LP4 and expected to begin qualified production by the end of 2026. Its customers include automotive, industrial equipment, and networking equipment makers, as well as the medical device and defense and aerospace sectors.
In other words, even for the same older-generation memory, the supply and demand picture differs by application and customer. Micron's expansion plan does not necessarily translate into better LPDDR4X procurement for low-priced smartphones. Likewise, the production cutback and end-of-life plans of major manufacturers shown in 2025 do not mean every manufacturer has left the older standard. Even so, a market structure in which products that need the older standard remain while new supply is hard to obtain explains why reused memory is gaining ground.
Sourcing History and Testing Shape the Quality of Reused DRAM
Even if reused DRAM offers an advantage in procurement, the reliability and lifespan of smartphones that use it is a separate matter. In the public summary of its October 30, 2026 report, TrendForce explains that reused memory comes from diverse sources with varying quality, and that suppliers sort and grade it. Even 4GB LPDDR4X parts of the same type do not necessarily share the same pre-recovery usage conditions or quality.
In its November 6 announcement, TrendForce further pointed out that complex sourcing routes and quality uncertainty could reduce product reliability and shorten smartphone lifespans. As a result, manufacturers could also face higher warranty and repair costs. However, this is an assessment of the risks associated with quality control of reused parts; it is not a measured confirmation that failure rates have risen or lifespans have shortened in any specific smartphone.
There are also points of caution in the processing itself. Regarding general BGA processing, Kei-All explains that when "underfill," a resin that reinforces the underside of a component, has been used, removing the resin is difficult and the risk of damaging the component rises. In Section 4 of a technical document on PBGA rework published by Texas Instruments, emphasis is placed on not overheating surrounding components during removal and on properly controlling temperature and placement accuracy. It also does not recommend manual placement that relies on visual alignment alone.
However, these are general materials on BGA and PBGA processing and do not show that all the reused DRAM in question has the same problems. Nor are they results from measuring the failure rate or remaining lifespan of smartphones carrying reused LPDDR4X. The question of what controls are needed during processing and the question of how long the finished product can actually be used must be evaluated separately.
| Item to check | What public materials show | What cannot be judged from that alone |
|---|---|---|
| Standard/capacity | Current adoption centers on 4GB LPDDR4X parts | Source of recovery and past usage history |
| Processing method | Reballing is a process of reattaching solder balls for connection | Whether lifespan is equivalent to new parts |
| Product quality | Besides sourcing, the content of sorting and testing matters | Failure rates of individual models or lifespan differences from new parts |
A memory's standard and capacity indicate what specifications the component has. The description of the reballing process, meanwhile, shows only how the connection to the board is rebuilt. Judging actual product quality requires checking the component's sourcing history, sorting criteria, and the content of the tests performed. Without distinguishing these, one would end up concluding, without sufficient basis, either that "used means failure-prone" or that "re-soldered means as good as new."
Notably, TrendForce's public November 6 announcement does not name the manufacturers adopting reused DRAM or specific models, nor does it give failure rates or lifespan comparison data against new parts. This is limited to what can be confirmed from the body of that day's public press release and does not mean that no industry-wide test data exists. Determining whether a specific smartphone sold in Japan uses reused DRAM, and how reliable that product is compared with devices using new DRAM, would require model-by-model information.
How Far Can Reused Memory Hold Down DRAM Price Surges?
TrendForce forecasts that DRAM price increases will accelerate in January–March 2027, and that adoption of reused memory could spread to smartphones in higher price ranges than today. This is an outlook for future market trends, not a record of adoption already progressing in high-priced devices. Even so, for device makers that must keep using LPDDR4X, recovered parts could serve as an alternative to new DRAM.
However, adopting reused memory does not necessarily guarantee a fixed level of cost savings. According to the public summary of TrendForce's October 30 report, reused memory differs in source and quality and is sold after sorting and grading, so prices also vary widely. Even if the volume of recovered parts is large, not all of them are suitable for reuse in smartphones, and the procurement price of parts that can actually be adopted must be considered separately. In addition, the November 6 announcement gives no specific cost-reduction rate, and it is also unclear how much any suppression of parts prices would be reflected in smartphone retail prices.
The forecast that adoption will reach 7–9% in 2027 also does not mean the DRAM shortage will ease by the same proportion. What TrendForce presented is the adoption rate in the smartphone market, not the share of "bit consumption," the total capacity of mobile DRAM overall. Given that current adoption is concentrated mainly in 4GB parts, the share of devices carrying reused memory must be distinguished from its contribution to total DRAM demand. Moreover, because the detailed estimation method has not been disclosed, this adoption rate cannot simply be converted into smartphone shipment volumes.
In its November 6 announcement, TrendForce forecasts that even if adoption of reused memory expands in 2027, its share of annual bit consumption of smartphone LPDRAM will remain relatively small. The public summary of the October 30 report likewise assesses the impact on overall DRAM bit demand as limited and says it is unlikely that the spread of reused parts will significantly change the upward trend in DRAM contract prices. The view, in short, is that while reused memory may become a procurement option for individual device makers, it is unlikely to be a decisive means of improving supply and demand in the DRAM market as a whole.
For consumers, what matters is how the adoption of reused memory affects not only product prices but also quality and warranties. The memory capacity and standard listed on a smartphone's spec sheet do not reveal whether the part is new or recovered, or what tests it has undergone. Going forward, how much information manufacturers disclose about parts sourcing and quality control, and what warranties they attach to finished products, will become a basis for judgment. If that transparency and quality control are ensured, reused DRAM could become a realistic option for sustaining the supply of smartphones that need older-generation memory.