Memory Card Prices Double in Four Months: How Long Will the Surge Last?

marsbitPublished on 2026-04-16Last updated on 2026-04-16

Abstract

NAND flash memory prices have entered a rapid upward cycle, with consumer-grade storage products like microSD cards seeing significant retail price increases. For example, a SanDisk Extreme 128GB microSD card rose from $17 in October 2025 to nearly $40 by February 2026—a 130% surge in under four months. This price surge is driven by structural shifts in the NAND market, primarily due to soaring demand from AI data centers. These large-scale buyers are securing the majority of NAND wafer supply through long-term contracts, leaving limited inventory for the consumer market. According to TrendForce, NAND contract prices rose 55–60% in Q1 2026, with enterprise SSD prices climbing 53–58%. Retail prices rose even more sharply due to constrained supply in the distribution channel. Unlike the 2016–2017 price cycle caused by production transitions, the current spike is demand-led. AI data centers are consuming NAND capacity at an unprecedented rate, with 2026 demand growth estimated at 20–22% against supply growth of only 15–17%. Manufacturers are prioritizing high-margin enterprise products over consumer-grade storage, further tightening retail availability. New production capacity from major suppliers like Samsung, Micron, and Kioxia is not expected until late 2027 or 2028. Until then, consumer storage prices are likely to remain high, with no significant price relief anticipated in the near term.

Since the beginning of this year, NAND flash prices have entered a new round of rapid increases, with retail prices of consumer storage products being the first to be affected. In October 2025, a SanDisk Extreme 128GB microSD card was sold for $17 on Amazon. By February of this year, the price of the same card had risen to nearly $40. In less than four months, the price increased by 130%.

First, it is important to clarify the difference between memory modules (RAM) and memory cards, as they are not the same type of product. Memory modules (RAM) are temporary storage inside computers, used for reading and writing data when programs are running; data is lost when power is cut off. Memory cards (such as microSD cards) are external expandable storage used for long-term preservation of files like photos and videos; data is not lost when power is cut off. The price increase discussed in this article refers to the latter—memory cards and the underlying NAND flash chips.

The continuous rise in memory card prices is driven by a systemic repricing of the entire NAND flash market. The starting point of this repricing is AI data centers competing for the same wafers.

Contract Prices Rose 50%, but Retail Prices Rose 130% by the Time They Reach You

First, let’s discuss what is happening.

Global NAND flash contract prices began climbing rapidly at the end of last year. According to a February report by market research firm TrendForce, overall NAND contract prices in the first quarter of 2026 increased by approximately 55–60% compared to the fourth quarter of last year, with enterprise SSD prices rising by 53–58%, setting a new record for quarterly growth. TrendForce also predicted that overall NAND contract prices would rise by another 70–75% in the second quarter.

These numbers represent the bulk contract unit prices agreed upon between large customers and do not directly equate to the retail prices on e-commerce platforms. However, retail prices on the consumer side have risen even more sharply than contract prices. The rightmost bar in Figure 1, showing a 130% increase, represents the actual price shock felt by ordinary consumers.

Why are retail price increases far exceeding contract price increases? Because the consumer market is a "residual allocation market." When NAND manufacturers set delivery plans, they prioritize fulfilling orders for large customers with long-term framework agreements, including AI data center operators and hyperscale cloud service providers. Only after these orders are fulfilled do the remaining inventories enter the distribution channels of the consumer market. With supply compressed, the spot market has almost no buffer capacity against price increases, leading to steeper retail price hikes than contract price hikes.

Kingston has publicly confirmed that its procurement cost for NAND wafers has increased by 246% compared to a year ago. This is a cost shock at the raw material level, which is ultimately passed down to consumers through product prices.

How AI Is Pushing Up the Price of a Memory Card

There are two key points in this graph worth highlighting.

The first is around October 2025, when relatively low-priced memory cards were available on the market. That period marked the end of the previous cycle of oversupply. From 2023 to 2024, major storage manufacturers accumulated large inventories amid weakening demand, leading to continuously declining prices. Photographers, creators, and gamers took advantage of this window to stock up on memory cards at historically low prices.

The second key point is the fourth quarter of 2025. Samsung, Kioxia, Micron, and SK Hynix successively announced production cuts and price increases, completely reversing the situation in a short period. Samsung raised prices for enterprise customers by over 100%, while Kioxia explicitly stated that its entire 2026 production capacity had been pre-sold to large customers, effectively cutting off supply to the consumer market.

Since then, retail prices for memory cards have been climbing steadily and are expected to reach the $50–$60 range by mid-2026, with no window for a price correction throughout the year. This is not market speculation but a structural adjustment in the supply allocation mechanism. Before AI data centers became the top-priority buyers in the NAND market, consumer and enterprise products participated more or less equally in capacity allocation. Now, the consumer end is at the very bottom of the allocation chain.

This Time, It’s Completely Different from 2017

The NAND industry experiences a price cycle approximately every three to four years. The last significant price increase occurred in 2016–2017 and lasted nearly two years. That round was triggered by the technological transition from 2D NAND to 3D NAND. The new stacking process slowed effective output during the yield ramp-up phase, tightening supply and driving up costs. However, once 3D NAND production lines stabilized for various manufacturers, Samsung, SK Hynix, and Micron simultaneously expanded production significantly. Inventories quickly shifted from shortage to surplus, and prices plummeted in early 2018.

This time, the driving force is entirely different, and the path to recovery will also be截然不同.

According to TrendForce data, global NAND demand growth is expected to reach 20–22% in 2026, while supply-side growth is only 15–17%. The absolute gap is not large, but in a market of this scale, a few percentage points of supply-demand disparity can trigger extremely sharp price reactions. More importantly, this gap is not caused by technical issues but by a structural shift in demand. AI data centers are consuming NAND capacity in a sustained, massive, and high-priority manner, and the scale of this demand has no ceiling.

New production capacity to alleviate supply constraints will not arrive until late 2027 to 2028. Samsung’s NAND production lines at its P4 facility in Pyeongtaek, Gyeonggi Province; Micron’s new wafer fab in Idaho, USA; and Kioxia’s expansion of its Iwate factory all point to this timeframe. 2026 is the year with the largest supply-demand gap, not the turning point for prices.

Manufacturers Aren’t Lacking Capacity; They’re Actively Selling It to the Highest Bidder

The following graph illustrates the essential mechanism behind this price surge. In the NAND industry’s revenue structure, the share of enterprise products (AI data center SSDs, general server storage) is rapidly expanding. According to comprehensive estimates by industry agencies, the share of enterprise products in overall NAND revenue has risen from about 45% in 2024 to about 62% in 2026, while the share of the consumer and client market has been compressed from 55% to about 38%.

The logic driving this shift is straightforward: for the same wafer area, producing enterprise-grade high-density QLC SSDs yields 3–5 times the unit profit compared to producing consumer-grade memory cards. Capacity allocation by manufacturers like Kioxia and Samsung follows the principle of maximizing commercial利益, allocating the best wafers to the highest bidders.

This mechanism also has a hidden effect. As available inventory in the consumer market decreases, distributors and retailers accelerate their stocking pace to hedge against future price increases, further accelerating inventory drawdown on the consumer end and creating a self-reinforcing cycle of price hikes.

For consumers, memory card prices will remain high for a considerable period. This is not due to a lack of wafer capacity but because the allocation priority of the consumer market has been systematically lowered. Only when the construction pace of AI computing infrastructure slows down will excess wafer capacity return to the consumer product supply chain, but that won’t happen until after 2027.

The SD card in your camera and the world’s largest AI data center are using the same wafer. Now you know who wins.

Related Questions

QWhat is the main reason for the sharp price increase in memory cards over the past four months?

AThe price surge is primarily driven by structural demand shifts, with AI data centers consuming NAND flash capacity at a high priority, leaving limited supply for the retail consumer market.

QHow does the retail price increase of memory cards compare to the NAND flash contract price rise?

AThe retail price increased by about 130%, which is significantly higher than the contract price rise of 55-60% in Q1 2026, due to the consumer market being a 'residual allocation market' with minimal buffer against supply constraints.

QWhy is the current NAND price cycle different from the one in 2016-2017?

AThe 2016-2017 cycle was caused by technical transitions (2D to 3D NAND) and resolved with improved yields and expanded production, while the current cycle is driven by structural AI demand, with no quick supply fix until 2027-2028.

QWhat role do AI data centers play in the NAND market supply chain?

AAI data centers are the highest-priority buyers, securing most of the NAND产能 through long-term contracts, which reduces the allocation for consumer products like memory cards and pushes retail prices higher.

QWhen is the NAND supply shortage expected to ease, and why?

AThe shortage is expected to ease around late 2027 to 2028, when new产能 expansions from major manufacturers like Samsung, Micron, and Kioxia become operational, increasing overall supply.

Related Reads

Super-Rich Hoarded Record Cash in February, Stock Market Hit New Highs Four Months Later: Who's Getting Fooled?

In February, the total assets in US money market funds reached a record high of approximately $8.25 trillion, a trend highlighted by high-net-worth individuals increasing their cash holdings. Notably, Warren Buffett's Berkshire Hathaway amassed a $381.7 billion cash pile ahead of his 2025 retirement, while other prominent figures like Peter Thiel sold tech stocks, fueling narratives of wealthy investors seeking safety. However, by June, the trend reversed. Money market fund assets fell to around $7.87 trillion, indicating a flow of capital back into equities. Concurrently, the S&P 500 and Nasdaq reached all-time highs, with the S&P 500 surpassing 7600 points. This market surge occurred despite the earlier defensive moves, highlighting a potential opportunity cost for those who retreated to cash. Analysis shows that since early 2022, the S&P 500's total return significantly outpaced that of prime money market funds. The capital shifted from equities appears to have been partly reallocated into alternative investments like real estate, art, and private credit, especially among ultra-high-net-worth individuals. Meanwhile, major investment banks like Goldman Sachs and Morgan Stanley have raised their year-end targets for the S&P 500, citing AI-driven earnings growth, while also cautioning about risks including market concentration and economic fragility beneath the surface rally.

marsbit14m ago

Super-Rich Hoarded Record Cash in February, Stock Market Hit New Highs Four Months Later: Who's Getting Fooled?

marsbit14m ago

Robot Vacuums Have Been Competing for 20 Years, So Why Are 90% of Chinese Households Still Hesitant?

The article explores why over 90% of Chinese households are still hesitant to adopt robotic vacuum cleaners despite two decades of industry development, identifying a core "trust gap" as the primary barrier. The central issue is not a lack of need, but user concerns about reliability in dynamic, real-world home environments. Common anxieties include the robot dragging pet waste, colliding with transparent objects, tangling in cords, scattering cat litter, getting stuck on thresholds, missing corners under furniture, and requiring high-maintenance bases that develop odors. The industry's past focus on competing on technical specs (suction power, mopping functions) has not adequately addressed these practical usability and trust problems. The piece then examines DJI's entry into the market with its ROMO 2 model as a potential new approach. Leveraging its expertise in spatial perception and obstacle avoidance from drones, DJI's solution emphasizes "less intervention" through three key principles: less manual re-cleaning, less user rescue missions, and less maintenance. Specific ROMO 2 features highlighted include advanced obstacle recognition (handling transparent objects and small items), adaptive leg mechanisms for climbing thresholds (up to 8.5cm), an extendable arm for reaching under furniture, AI for identifying and appropriately handling different mess types (e.g., avoiding scattering dry debris), and a self-cleaning base designed to minimize user upkeep. The article argues the next phase of competition should shift from a "parameter race" to a "trust race." It draws a parallel to the iPhone's simplification of the smartphone, suggesting that focusing on a reliable, low-hassle user experience—where people feel confident leaving their floors to the machine—is what's needed to finally convince the vast majority of观望ing families. The ultimate test for products like the ROMO 2 will be long-term user adoption, retention, and口碑, not just technical specifications.

marsbit14m ago

Robot Vacuums Have Been Competing for 20 Years, So Why Are 90% of Chinese Households Still Hesitant?

marsbit14m ago

The Unclear American Economy: Resilient or Cooling Down?

**U.S. Economic Outlook: Resilient or Cooling Down?** This analysis examines whether the U.S. economy is heading towards a recession. While still growing, the economy shows significant signs of strain. Key data points include Q1 2026 GDP growth of 1.6% and Q1 PCE inflation at 4.5% (annualized), more than double the Fed's target. The labor market remains resilient but is softening, with unemployment at 4.3%. Critical recession indicators present a mixed picture: the yield curve has normalized after a prolonged inversion (historically a late-cycle signal), and the Conference Board's Leading Economic Index has been declining. Current recession probability for 2026 is estimated at 19%, but rises to 41% for 2027, indicating heightened delayed risks. Major pressures are building: a wall of corporate debt refinancing at higher rates, depleted consumer savings, a contracting housing sector, and an energy price shock. The economy exhibits stagflationary characteristics—high inflation alongside slowing growth—which constrains the Federal Reserve's policy options. Historical patterns show recessions are often preceded by Fed tightening and yield curve inversions. If a recession occurs, it is expected to be mild, similar to 2001 rather than 2008. For investors, a defensive portfolio shift toward staples, healthcare, and short-term high-quality bonds may be prudent, while maintaining a long-term, diversified perspective. Key developments to monitor include upcoming GDP, employment, and inflation data, as well as policy signals from the new Fed Chair.

marsbit21m ago

The Unclear American Economy: Resilient or Cooling Down?

marsbit21m ago

The Most Advanced Large Models Are Now Subject to Export Controls Like Enriched Uranium

In an unprecedented move mirroring the control of enriched uranium, the US Commerce Department has imposed an export control ban on Anthropic's advanced AI models, Fable 5 and Mythos 5, forcing their global shutdown. This marks the first time a purely digital entity—a set of neural network weights—has been subjected to such hardware-like strategic export restrictions, based not on physical scarcity but on its concentrated "capability density." The article draws a direct parallel to the historical control of nuclear technology, arguing that just as uranium ore becomes a controlled substance only when enriched to a critical threshold, AI capabilities become subject to regulation when compressed into a single, potent, and easily accessible interface. This "enriched AI" is seen as crossing a threshold where its aggregated power poses a potential threat. The author predicts three major consequences over the next decade. First, capability auditing will become institutionalized, with governments setting compliance checklists and thresholds for model power, triggering automatic export controls. Second, jurisdictional boundaries will blur as US export controls extend their reach globally, governing any user of American AI services regardless of location, forcing non-US entities to reconsider their AI supply chain dependencies. Third, a technological bifurcation will occur, splitting the AI landscape into a restricted, high-risk track of advanced US proprietary models and a more reliable track of open-source or locally developed alternatives, where guaranteed access may outweigh raw performance. The core crisis exposed is the lack of a legal property rights framework for AI "intelligence." While companies invest heavily in integrating these models into their production systems, legally they only purchase a service that can be revoked at any time, leaving them with no recourse for their sunk investments. The conclusion warns of a permanently fractured digital world where the most capable models may not be the most usable, and clear, unassailable ownership of technology will become paramount.

marsbit34m ago

The Most Advanced Large Models Are Now Subject to Export Controls Like Enriched Uranium

marsbit34m ago

Trading

Spot
Futures
活动图片