"When will the 'storage frenzy' peak? This is the most effective 'leading indicator.'"
```
Since January 2023, driven by AI computing demand, storage chip stocks have continued to surge, with the share prices of the top three storage manufacturers averaging an increase of 699%. However, as stock prices keep rising, the market's main concern is no longer "how much more can they rise," but "what signals to watch for in order to judge the turning point."
On March 16, the UBS Global Research team released a report, "Global I/O Storage Semiconductors," reviewing the cyclical patterns of the storage industry over the past 20 years, and reassessing current leading indicators. UBS pointed out that AI computing has fundamentally changed the underlying logic of the storage industry, and traditional valuation and forecasting models may no longer be applicable. Operating profit has become a better leading indicator.
The Industry's "Underlying Logic" Has Changed: AI Pushes Supply-Demand Relations to a New Equilibrium
UBS attributes the key factor of this round of market performance to "the rise in storage value in the AI computing era." The report points out that behind the revaluation of storage, two supply-side constraints are accumulating:
- HBM is occupying more and more DRAM wafer capacity, resulting in "severe DRAM shortage";
- This tension is further amplified by the "trade ratio": the die size of HBM DRAM continues to increase relative to DDR, so the unit HBM "consumes" more capacity.
On this basis, UBS puts forward a conclusion that is more sensitive for investors—the return rate center has shifted upwards. The report bluntly says: "We believe ROE has been structurally reset," and estimates that the average ROE for Samsung, SK Hynix, and Micron from 2026 to 2030 will be 36%, significantly higher than the previous decade's 15%. This means: using the old cycle template to find the market top might fail more frequently.
Traditional Indicators Are No Longer Reliable: The "Second Derivative" Fails
In the past, investors often preferred to use the "second derivative"—that is, the quarter in which the acceleration of storage contract prices (ASP) was fastest quarter-on-quarter or year-on-year—to predict the stock price top. However, UBS's review shows that the reliability of this indicator is declining.
Out of 10 "stock price tops" in the past 20 years, only 50% of the time did stock prices peak in the same quarter or neighboring quarter as the quarter-on-quarter change in DRAM ASP.
For example, in Q4 2009 (post-global financial crisis recovery), Q2 2013 (post-industry consolidation cycle), and Q1 2017 (traditional cycle), the peak in ASP quarter-on-quarter change occurred 3, 5, and 5 quarters before the stock price peak, respectively.
UBS points out that though actual ASP and stock prices are slightly more synchronous (60% peaking in the same quarter), overall, the "second derivative" is no longer a precise "escape" signal in the current complex market environment.
Looking for a New Anchor: Operating Profit Is a Better Leading Indicator
Since traditional indicators have failed, what should investors watch? UBS's answer is: Operating Profit (OP).
Operating profit not only reflects price changes, but also integrates factors such as bit growth and cost reduction per bit. Therefore, it is closer to the industry's true prosperity "end value."
Report analysis shows that, in the past 20 years, stock prices peaked at the same time or before operating profit in 90% of cases.
Especially before 2012, stock prices and operating profits peaked almost simultaneously. Since then, the stock market has become more anticipatory, with stock prices usually peaking one to two quarters before operating profit (most cases, one quarter earlier).
However, UBS also reminds investors that predicting when operating profit will peak is not easy. The reason is still that structural changes in supply and demand brought by AI may make profit rhythms harder to predict, especially as HBM continues to crowd out DRAM capacity. Price, supply, and profit linkages become more complex, and the "estimated timing" of profit peaks may quickly drift.
Therefore, operating profit can be used as an important observation indicator, but it is by no means a "panacea."
AI Reshapes the Industry: Structural Reset of ROE, Bull Run May Continue to 2027
UBS emphasizes that the current storage cycle is fundamentally different from the past. The advent of the AI computing era has caused a fundamental shift of value toward storage.
As HBM (High Bandwidth Memory) occupies more and more DRAM wafer capacity, the DRAM shortage has become increasingly severe. Moreover, the chip size of HBM DRAM keeps increasing, further aggravating capacity strains.
Based on these factors, UBS believes the net asset return rate (ROE) of the storage industry has undergone a structural reset. The report predicts that from 2026 to 2030, the average ROE of Samsung, SK Hynix, and Micron will reach 36%, far higher than the past decade's 15%.
Therefore, the report remains optimistic about the outlook for storage stocks. The report predicts that storage industry operating profit will peak in Q3 2027. If other conditions remain unchanged, this means the bull run for storage stocks could continue until Q2 2027.
UBS continues to favor SK Hynix as its top buy, and maintains a "buy" rating on Samsung, Micron (MU), and Nanya Technology.
~~~~~~~~~~~~~~~~~~~~~~~~
The above wonderful content comes from Chasing Wind Trading Desk.
For more detailed interpretation, including real-time analysis and frontline research, please join [Chasing Wind Trading Desk Annual Membership]
Risk Warning and DisclaimerThe market has risks, and investment must be cautious. This article does not constitute personal investment advice, nor does it take into account the special investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular situation. Invest accordingly, at your own responsibility. ```