Memory module prices plummeting, memory stock prices falling across the board, has the memory super cycle peaked?

Author: Long Yue

Recently, the memory prices, which have been rising for several months, suddenly reversed and dropped, raising concerns in the market about the peak of the memory cycle.

According to market tracking data, several retailers in the United States have significantly reduced the prices of DDR5 memory, with the highest single set drop reaching $100. Taking Corsair’s VENGEANCE series as an example, the current price of its 32GB capacity model with a maximum frequency of 6400MHz is about $379.99, a substantial decrease from the recent peak of $490, with a single set drop of over $110.

The domestic market has also been impacted, with wholesalers telling the “China Business Journal” that the price of mainstream 16GB memory sticks “dropped more than a hundred yuan in a day,” and large stockpilers are frantically selling off their inventory.

“It collapsed directly from last Saturday.” Boss Wang, a wholesaler who has been operating memory devices in Baidu Brain for many years, admitted to the media. He showed an extreme price curve of a mainstream 16GB 3200MHz memory stick: it was just over 130 yuan last May, then skyrocketed to a peak of 980 yuan in December, but after several months of high fluctuations, the current spot price has fallen back to around 700 yuan.

Boss Wang lamented that due to the price increase exhausting consumer expectations, “if it’s not a necessity, people won’t buy it. Compared to before November last year, our sales have dropped by more than 60%.”

At the same time, Google released a paper on a new compression algorithm called “TurboQuant.” The research pointed out that this technology can reduce the memory usage of key-value caches (KV Cache) during the operation of large language models by at least 60%. Investors quickly priced this in: the AI hardware shortage issue will fundamentally ease, and memory demand will be significantly reduced.

The chill in the spot market quickly transmitted to the capital market. Micron Technology’s stock price has retreated more than 24% from its recent peak, and Western Digital has also dropped nearly 21% from its high of $777.60. Meanwhile, last week, the market value of the U.S. memory chip sector evaporated by nearly $100 billion.

Faced with the price plunge and stock price crash, market participants have developed serious divisions regarding the future of the memory industry. Some investors believe that the traditional “pig cycle” of memory has peaked, while HSBC believes that market concerns are excessive and that we are currently in the middle of an AI-driven memory supercycle, with strong demand for high-end products like HBM, and the memory shortage may last for one to two years.

Buyers say “no”: Is the traditional “pig cycle” repeating?

For traders who follow traditional cycles, the market crash is not that simple. Former journalist and renowned semiconductor analyst Dan Nystedt pointed out that many bulls attribute the recent plunge to Google’s paper, but this is just the surface. Dan believes the real reason is that the prices of some smartphone memory chips have stopped rising.

“The real reason is much simpler: the prices of certain smartphone memory chips have stopped rising. Buyers ultimately said ‘no’—this is the first peak signal that experienced memory cycle investors look for before selling.”

Dan Nystedt stated that due to excessively high DRAM and NAND prices, some smartphone manufacturers plan to reduce or even cancel production of mid-range and low-end phones in 2026. He revealed that two weeks ago, buyers rejected higher DDR4 prices.

Dan Nystedt likened the memory industry to the “pig cycle” in agriculture: high prices encourage enterprises to expand production, but building factories takes time, and when new capacity is released simultaneously, prices will crash. He believes that investors following this script have quickly exited, causing significant corrections in the stock prices of Micron and SanDisk.

Over the past 50 years, memory chips have undergone more than a dozen major boom/bust cycles. Since 2010 alone, there have been three: the 3G/4G and cloud computing boom from 2012 to 2015; the 5G and cloud service provider expansion from 2016 to 2019; and the pandemic-driven surge in PC/server from 2020 to 2023. The cycle that began in 2024 is driven by AI servers (HBM and SRAM).

“Whenever someone writes ‘this time is different,’ it is usually a classic sign of bullish sentiment going crazy.” Nystedt quoted the legendary trader Jesse Livermore: “The market is always right, while opinions are often wrong.” He reminded investors that when chip buyers no longer panic buy, and rebounds repeatedly encounter sustained selling, seasoned funds will quickly withdraw according to the script.

Structural Transformation: Memory Companies Are No Longer “Cyclical Stocks”?

However, independent analyst Jukan has a different view on Dan Nystedt’s analysis.

He pointed out that buyer resistance to price increases is mainly focused on traditional memory like DDR4, not the entire memory market. The previous abnormal surge in DDR4 prices was partly attributed to stockpiling in the Chinese market, which gave smartphone manufacturers room to adjust the specifications of low-end devices.

“But DDR5 is a completely different matter.” Jukan pointed out that smartphone and PC manufacturers in the first and even second quarters of this year have honestly accepted the significant price increases of DDR5. In the current AI and high-end device ecosystem, DDR5 is not a negotiable bargaining chip; rather, it is a core input that buyers must ensure even if they have to pay a premium. Flagship products built around DDR5 simply cannot reduce their specifications.

Secondly, the market has completely ignored the fundamental transformation of the business models of memory giants. Jukan scoffed at the so-called “seasoned investors” who blindly sell off based on spot price declines.

“The way memory companies operate is no longer the blind expansion model of the past.” Jukan keenly pointed out that the three giants—Samsung, SK Hynix, and Micron—are moving toward TSMC’s business model—only building capacity after securing advance payments from core customers and long-term visibility of demand.

Recently, South Korean media reported that Samsung is discussing prepaid cooperation agreements with giants like Microsoft. Memory giants are more aware than anyone of the pain caused by overcapacity destroying cycles. Therefore, they are now pursuing extremely restrained capacity expansion rather than reflexive overbuilding.

Investment Banks Support: The Memory Supercycle Is Only at Midpoint, Market’s Five Major Concerns Are Overblown

In contrast to the panic in the spot market, investment banks remain confident about the long-term prospects of the memory industry. HSBC clearly stated in a research report released on March 30, “In our view, the current concerns are overblown; we are at the midpoint of an AI-driven supercycle.”

The current market concerns are all excessive reactions, and the bank listed five specific worries:

1. Negative impacts of rising raw material and electricity prices due to conflicts in the Middle East;

2. Slowdown in memory price growth in the second half of 2026;

3. Industry technologies such as Google’s “TurboQuant” and NVIDIA’s “KVTC” that reduce memory usage in AI systems;

4. Gradually increasing capital expenditure plans from major memory manufacturers;

5. Intensified competition from Chinese memory manufacturers.

The report pointed out that the Middle East conflict does not have a substantial impact on the procurement of raw materials by memory manufacturers. Furthermore, the absolute growth of profits will have a much greater impact on stock prices than the slowdown in the slope of DRAM price increases. At the same time, memory manufacturers remain highly aware and restrained in executing their capital expenditures.

Regarding Google’s TurboQuant technology, which triggered market sell-offs, the bank believes it is still too early to worry. The commercialization of this technology will take about a year, and its reference parameters are smaller than the current AI environment. More importantly, the bank pointed out that TurboQuant alleviates memory bandwidth bottlenecks, improving system efficiency and reducing token costs, thereby accelerating the commercialization and popularization of AI. The report stated:

“The net effect is that we believe the increase in efficiency will accelerate AI development—this is a positive event that should trigger a sharp increase in AI adoption rates.”

At the same time, the bank expects that AI server shipments will increase by 28% year-on-year in 2026. From 2026 to 2027, the average DRAM content per server will achieve a robust growth of 17%. As the demand for AI inference explodes, enterprise solid-state drives (eSSD) are entering a golden age. The report predicts that by 2027, the proportion of eSSD in total NAND demand will soar from 18% in 2023 to 40%. AI servers will consume 62% of that.

The bank believes that the current market is at the midpoint of an AI-driven supercycle, comparable in scale to the six-year DRAM shortage caused by office automation from 1990 to 1995. Looking back at history, from 1990 to 1995, along with the popularization of Windows 3.0 and subsequent operating systems, office automation triggered a structural shortage of DRAM that lasted six years, driving the DRAM market size from $7 billion in 1990 to a staggering $41 billion in 1995.

The bank believes that the current infrastructure construction driven by large models, agentic AI, and physical AI (such as autonomous driving) will result in a memory shortage that will last at least one to two years.

Based on these judgments, the report firmly views the certainty of their benefits in the memory supercycle. Regarding the recent plunge, the report stated: “We believe any pullback provides an additional buying opportunity.”