A warning from one of the storage industry’s most connected insiders is raising major concerns. Khein-Seng Pua, CEO of Phison Electronics, a company that controls roughly 20% of the global SSD controller market, recently offered a stark assessment of where the memory and storage market is heading. In a recent interview, Pua suggested that the accelerating NAND flash shortage could force deep production cuts and push smaller companies to exit product lines or even go under in the second half of 2026. For the ITAD and secondary market, the implications deserve serious attention.
What is NAND flash, and why should ITADs care?
For readers less familiar with the component side of the devices processed every day in this industry, a quick primer. NAND flash is the semiconductor memory technology that powers solid-state drives (SSDs) and the embedded storage chips found inside laptops, tablets, smartphones, and servers. Unlike traditional spinning hard disk drives (HDDs), which store data on magnetic platters, flash memory has no moving parts, operates faster, and has become the dominant storage format in most modern enterprise and consumer devices. Most laptops, tablets, and smartphones manufactured after roughly 2015 rely on NAND flash storage, though budget devices, enterprise storage appliances, and high-capacity NAS systems continued using traditional hard drives well beyond that date.
The NAND market is supplied by a small number of manufacturers: Samsung, SK Hynix, Micron, Kioxia, and Western Digital chief among them. When their supply tightens, the effects move through every industry that touches electronics.
What is driving the shortage?
The primary driver is AI infrastructure demand. Data centers building out the compute and storage capacity needed to train and run large AI models are consuming NAND flash at an unprecedented rate. Supply that would normally flow to consumer electronics, automotive, and enterprise PC manufacturers is increasingly being absorbed at the top of the market by hyperscalers willing to pay premium prices.
Pua described witnessing a meeting in China where major mobile and automotive companies were essentially pleading with suppliers for flash memory allocation and coming up short. He suggested that smaller firms may struggle to secure meaningful supply at all in the months ahead.
On pricing, Pua cited sharp component cost increases as evidence of how distorted the market has become, though specific figures he referenced have not been independently verified, and should be understood as reflecting spot market conditions in constrained categories rather than broad market averages. In that interview, he pointed to one example where 8 GB eMMC reportedly rose from about $1.50 to roughly $20 within a year, a change he described as emblematic of how extreme pricing has become and noted was still not enough to guarantee supply. What is clear from multiple industry sources is that NAND contract prices rose sharply through late 2025, with some suppliers reportedly lifting contract prices by around 50%, and that spot market pricing in some categories has moved well beyond recent norms.
Pua also claimed that at least one major NAND foundry has begun demanding multi-year cash prepayments from customers, a practice he described as unprecedented in his experience. This cannot be independently confirmed through public records as well but was presented as a firsthand account from someone with direct supplier relationships.
Western Digital has signaled extremely tight supply conditions across its hard drive portfolio, with its CEO indicating that the company is essentially sold out of HDD capacity for calendar 2026, largely due to multi-year commitments with top cloud customers. It is worth noting that hard disk drives use magnetic storage rather than NAND flash and face a related but distinct supply dynamic, driven more by data center demand for high-capacity spinning storage than by the semiconductor shortage affecting flash memory directly. The two trends are running in parallel and compounding each other, but they have different root causes.
The ITAD and secondary market implications
The effects on the ITAD industry and the secondary market are real, immediate, and cut in multiple directions. On the positive side, used devices containing flash storage carry higher residual component value today than they did 12 months ago. Laptops, tablets, and smartphones moving through processing facilities are worth more, particularly at the component level. For remarketing operations with the sophistication to evaluate whole-unit versus component-level value, this is a meaningful opportunity worth examining. A device that previously made economic sense to remarket as a whole unit may now be worth more when assessed for its storage components depending on condition, capacity, and specification.
Refurbishment economics are more complicated, however. Operations that replace failed or wiped storage as part of a grading and refurbishment process are paying more for replacement SSDs than they were a year ago. Margins on refurbished units tighten accordingly, and pricing models built on 2024 component costs are worth reassessing.
Perhaps most strategically significant for the longer term: if new device supply tightens as Pua projects, enterprise customers will extend refresh cycles and keep devices in service longer. Pua’s projections for production cuts in smartphones and PCs go well beyond the tone of mainstream analyst commentary as of early 2026 and are best understood as a worst-case scenario rather than a base-case outlook. Even a fraction of that scenario, however, represents a tailwind for lifecycle extension services, refurbishment programs, and redeployment offerings. The circular economy argument has always made environmental sense. Increasingly it is making supply chain sense as well.
Finally, the OEM circular programs are worth watching in this industry, including HP Renew, Dell Asset Disposition, and Lenovo’s recently announced lifecycle services,as they face the same component cost pressures as independent ITADs. The economics of refurbishment get harder for everyone when storage costs rise. Independent ITADs with flexible procurement strategies and diversified downstream channels may prove more flexible in the near term than OEM programs built on more rigid supply chain structures.
What to do right now
ITAD industry stakeholders should review component valuation models and buyback pricing as a priority, since residual value assumptions built on pre-2025 data are likely outdated. Grading and processing workflows are worth examining to determine whether they allow for component-level value assessment alongside whole-unit remarketing decisions. Engaging downstream buyers now on forward pricing visibility is advisable before market conditions shift further. And enterprise customers signaling extended refresh cycles deserve close attention, because those conversations represent an opening to position lifecycle services before OEM programs get there first.
Pua’s interview is one data point from one industry executive, and some of his projections are more aggressive than mainstream analyst consensus. But the underlying dynamic he is describing, an AI-driven reallocation of semiconductor supply away from the devices the ITAD industry processes and remarkets, is real and already affecting pricing.
