Click the blue text to follow us
1. The “Super Cycle” of Storage Chips in 2025: Why Focus on This Sector Now?
The storage chip industry in 2025 has already transcended the old framework of the “inventory cycle” and has been propelled into a new “super cycle” by the AI wave. On one side, giants like Samsung and SK Hynix are reducing traditional DDR4 production capacity by over 20%, fully committing to HBM (High Bandwidth Memory), which is core to AI computing power. On the other side, the demand for storage capacity and speed from AI large models and data centers is experiencing explosive growth. Market monitoring data indicates that the global storage chip market is expected to exceed historical highs in 2025. This combination of “precise contraction on the supply side + structural explosion on the demand side” has ushered in an upward cycle of “increased volume and price” for the sector. However, problems arise: many stocks have surged along with the industry boom, but are their prices a result of “overvaluation” or “value support”? At this time, the combination of “PB + gross margin” is the key weapon for investors to identify “value traps”.
2. Understanding Two Indicators: PB for “Asset Base” and Gross Margin for “Profitability”
To discuss investment value, one must first understand the industry’s fundamentals. The storage chip market in 2025 is no longer a traditional business of “selling memory cards”. First, on the demand side: AI large model training requires massive data storage. For instance, training a model with hundreds of billions of parameters requires tens of PB of storage capacity. Data centers must upgrade to HBM and other high-bandwidth storage to support AI inference. This “computing power demand” is not just an “incremental” change, but a “qualitative transformation”—since 2024, the demand growth for HBM has exceeded 100%, leading to a situation where “one chip is hard to find”. Then, on the supply side: giants like Samsung, SK Hynix, and Micron are collectively “reducing old capacity and increasing new capacity”—cutting over 20% of DDR4 production while shifting wafer fabs and equipment towards advanced processes like HBM and 3D NAND. This “supply contraction + structural upgrade” has directly caused storage chip prices to begin recovering from the fourth quarter of 2024, with storage prices in October 2025 rising over 30% compared to the beginning of the year. Under the dual influence of “AI-driven demand explosion + precise contraction on the supply side”, the storage chip industry has jumped from “overcapacity” directly into a “structural supply-demand imbalance” super cycle. However, the hotter the market, the more caution is needed—some stocks are rising due to “AI concept bubbles”, while others are rising due to “real value”. At this time, the combination of “PB + gross margin” is essential to distinguish between “overvaluation” and “value traps”.
3. The Combination Logic: Why PB + Gross Margin Can Avoid “Overvaluation”?
Using PB or gross margin alone can easily lead to pitfalls—they must be combined to filter out the “true value traps”. First, consider two counterexamples:
•Counterexample 1: Low PB + Low Gross Margin—A company has a PB of 1.5 (below the industry average of 4), but a gross margin of only 15% (below the industry average of 25%). It seems “cheap”, but is actually a “trap”: its “low PB” comes from a large amount of outdated DDR4 capacity, and the “low gross margin” indicates that this capacity cannot generate profit. The low stock price is “reasonable”, not a value trap.•Counterexample 2: High Gross Margin + High PB—A company has a gross margin of 35% (above the industry average), but a PB of 8 (far above the industry average of 4). This may be due to AI concept speculation, where the stock price has already priced in three years of growth. Once the industry’s heat declines, the stock price will quickly fall back, representing “overvaluation”. The true value trap: PB below the industry average + gross margin above the industry average. For example, in the third quarter of 2025, the average PB in the storage chip industry is 4, and the average gross margin is 25%. A company has a PB of 3 (below average) and a gross margin of 30% (above average). The significance of this data is:1.“Asset Base Underestimated”: Spending 3 to buy 1 of net assets is 1 less than the industry average; 2. “Stronger Profitability”: A gross margin 5 points higher than the industry average indicates it has advanced capacity (like HBM) or technological advantages, allowing it to earn more during the industry’s upward cycle. The stock price of such companies has “dual upward momentum”: one is “PB recovery” (the stock price approaches the industry average PB, e.g., rising from 3 to 4); the other is “profit growth” (high gross margin leads to increased net profit, further driving up the stock price)—this is the “true value” in the AI super cycle.
4. Practical Steps: Using PB + Gross Margin to Find “Value Traps”
Understanding the logic is not enough; practical steps are needed—four steps to find suitable targets. Step 1: Identify “Core Targets”—First, find the “hard players” in the storage chip industry chain: for example, those making HBM (Samsung, SK Hynix, with companies like Lanqi Technology producing HBM interface chips in the A-share market), those making advanced NAND (Yangtze Memory Technologies, with related supply chain companies like Jingce Electronics), those making storage controllers (Demingli), and those making NOR Flash (Zhaoyi Innovation). Avoid companies that are merely “riding the concept” (e.g., those only selling storage modules without core chip technology). Step 2: Check Latest Data—Go to Tonghuashun or Dongfang Caifu to check the “latest quarterly PB” and “deducted gross margin” (be sure to deduct non-recurring items to avoid interference from government subsidies, one-time orders, etc.). For example, in the third quarter of 2025, Demingli’s PB is 3.2 (below the industry average of 4), and its deducted gross margin is 28% (above the industry average of 25%), meeting the “trap” standard. Step 3: Verify “Asset Quality”—Check the company’s “capacity structure”: for example, if a company has a PB of 3, check how much of its net assets are “advanced capacity” (like HBM production lines, 3D NAND fabs) and how much is “outdated capacity” (DDR4). For instance, if Yangtze Memory Technologies has over 70% of its net assets in 3D NAND capacity, this part of the asset will see demand growth in the AI era, so a low PB is a “true undervaluation”. Step 4: Assess “Sustainability of Gross Margin”—Is the high gross margin “accidental” or “long-term”? For example, if a company’s quarterly gross margin suddenly rises to 30%, check whether it is due to a one-time HBM order or a long-term customer agreement (like binding orders with Nvidia or AMD). Only a “sustainable high gross margin” can support long-term stock price increases.
5. Four Pitfall Avoidance Points: Don’t Let Indicators Become “Traps”
Finally, here are four “avoidance points” to remind you not to let indicators become “traps”:
1.Don’t Ignore “Asset Quality”: The premise of low PB is that “net assets are useful”—if a company’s net assets consist of outdated DDR4 capacity, no matter how low the PB, it should not be touched; 2.Don’t Trust “One-Time High Gross Margin”: For example, if a company raises its gross margin to 35% due to a one-time military order but lacks long-term orders, this high gross margin is “fake”; 3.Adjust Standards According to Industry Cycle: 2025 is an industry upcycle, so PB can be relaxed (e.g., if the industry average is 4, 3 can be considered low); if the industry enters a down cycle, PB must be stricter (e.g., below 2); 4.Don’t Ignore “Industry Position”: For example, Samsung’s PB may be higher than the industry average (e.g., 5), but it is the leader in HBM with a gross margin of 45%. This high PB is a “reasonable premium”—because its technology and market share can support a higher valuation.
Disclaimer
This article is for investment logic analysis only and does not constitute any investment advice. The storage chip industry is influenced by multiple factors including macroeconomics, technological iteration, industry policies, and supply-demand relationships. Indicators such as PB and gross margin are for investment reference only and do not guarantee investment returns. Investors should make independent investment decisions based on their own risk tolerance. The market has risks, and investment requires caution.