Maintain Underperform: Short-term Strength but Long-term Valuation Too High, Target Price 40,000 JPY

Overview

Bernstein issued a research report on Kioxia Holdings (285A.JP). While acknowledging the earnings explosion and stock momentum brought by short-term NAND shortages, it points out through a Sum-of-the-Parts (SOTP) valuation model that the current stock price has excessively discounted future high-growth expectations. The report argues that with the rise of Chinese suppliers and global capacity recovery, the long-term profitability of the NAND industry will be suppressed to reasonable levels rather than the extremely high margins currently expected by the market. Therefore, the "Underperform" rating is maintained, and a 40,000 JPY target price is set, implying significant downside compared to the current price.

Core views

Divergence Between Short-term Momentum and Long-term Pressure: The report points out that the year-on-year Average Selling Price (ASP) growth of the NAND industry in Q2 2026 may reach as high as 60% quarter-over-quarter, and Kioxia's earnings will continue to grow significantly in the short term. However, this boom driven by extreme shortage is unsustainable. Institutions predict that after Q3 2026, the pace of price increases will slow down significantly; by H1 2027, NAND prices will reach cycle peaks, then begin to fall in H2 2027, and rapidly return to normal levels in 2028. Valuation Logic: SOTP Model Reveals 50% Overvaluation: Institutions use the Sum-of-the-Parts (SOTP) method for detailed valuation of Kioxia, splitting its business into three parts: 1. Short-term super profits (until 2028): Treated as temporary shortage dividends, valuing the excess cash returns generated to shareholders at 1x multiple. 2. Long-term normalized earnings (starting 2029): Assuming market equilibrium is reached in 2029, long-term Gross Margin (GM) returns to 35% (considered a reasonable estimate based on company guidance). Of this, 50% of shipments protected by Long-Term Agreements (LTA) receive 20x PE valuation, while the remaining 50% receives 10x PE valuation. 3. LTA breach deposit income (2029-2031): Expected that some LTAs will be forfeited by customers due to low market prices, allowing suppliers to acquire deposits, but this income is unsustainable, receiving only 3x PE valuation. Combining these three parts yields an intrinsic value per share of approximately 41,500 JPY, about 50% lower than the current share price. Core Contradiction: Long-term Gross Margin Assumption Too Optimistic: The current market pricing implies very high profitability expectations for Kioxia's long-term potential. Sensitivity analysis shows that to justify the current share price, Kioxia's long-term stable gross margin must reach around 65%, or LTA financial guarantee ratios must be as high as 80%, or shortage conditions must persist beyond 2032. The institutions believe any of these three scenarios are extremely unlikely. Historical data shows that even NAND industry cyclic peak gross margins are far below 65%, and normalized gross margins are usually in the 20%-30% range. Structural Risk: Impact of Chinese Manufacturers: The report specifically emphasizes the profound impact of the rise of Chinese storage manufacturers (especially YMTC) on the industry landscape. By 2028, YMTC is expected to become the third-largest NAND supplier globally. Intense competition will force prices downward, making it difficult for non-Chinese suppliers to maintain ultra-high profit margins. Additionally, aggressive price hikes strategies provide opportunities for Chinese manufacturers to enter the global supply chain (especially in consumer electronics), which may make memory supply a structural advantage for China in AI development, further squeezing overseas giants' premium space.

Analysis framework

The institution's analysis logic revolves mainly around "mean reversion after the cycle peak": 1. **Volume-Price Breakdown and Supply/Demand Deduction**: First, by tracking shipment volumes and NAND content penetration rates of PCs, smartphones, and servers, combined with capacity expansion plans of manufacturers like YMTC, deduce that supply and demand will rebalance by 2028, thereby judging price trends. 2. **LTA Mechanism Stress Test**: Instead of blindly believing Long-Term Agreements can lock in high prices, quantitatively analyze the relationship between "breach deposits" and "market price difference". When the market price is below the LTA base price by a certain margin (e.g., 16%-30%), it is economically more cost-effective for customers to choose forfeiture of deposits and purchase in the spot market, proving that LTA cannot prevent the arrival of the price cycle. 3. **Multi-Scenario SOTP Valuation**: Abandon single PE/PB valuation because current earnings are periodic outliers. By assigning different valuation multiples separately to "one-time cash flow", "sustainable core business earnings", and "occasional penalty income", more accurately reflect the true value of cash flows of different natures. 4. **Qualitative+Quantitative Verification of Competitive Landscape**: Introduce variables of rising market share in China, combined with historical gross margin centrals, to demonstrate the impracticality of a 65% long-term gross margin.

Methodology notes

• Valuation MethodSOTP Segment Valuation

  Sum-of-the-Parts Valuation

  When a company owns business segments or cash flow sources of vastly different natures (such as short-term superprofit cash vs. long-term normalized profits in this case), using a single Price-to-Earnings ratio is often misleading. The SOTP method requires analysts to value each business segment separately before summing up. In this article, the institution assigns 1x, 20x, and 10x multiples respectively to Kioxia's "Short-term shortage dividend", "Long-term LTA protected profits", and "Non-LTA profits" to derive a more rational intrinsic value.

• Industry/Industrial Analysis FrameworkSupply-demand framework

  Semiconductor Inventory and Capacity Cycle

  Semiconductors are a typical strong cyclical industry, where prices are driven by short-term supply-demand mismatches. The research report judges that 2028 will be the inflection point for supply-demand balance by tracking downstream terminal (PC/Phone) demand contraction signs and upstream (YMTC/Kioxia, etc.) capacity expansion rhythms, thereby anticipating prices will shift from soaring to crashing. Understanding this transmission chain is key to grasping semiconductor stock investment timing.

• Competition and Strategy FrameworkMoat / competitive advantage

  Impact of Geopolitics and Technology Diffusion on Pricing Power

  The report points out that despite export controls, the technological progress and capacity expansion of Chinese storage manufacturers are weakening the pricing power of overseas giants. When competitors can provide sufficiently cheap alternatives, the original leader's "moat" (i.e., ability to maintain high gross margins) gets filled flat. This reminds investors that when evaluating cyclical stocks with strong pricing power, potential new entrants (especially local rivals supported by policy) must be included in long-term margin assumptions.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Kioxia Holdings (285A.JP)

  Direct target. Report believes it benefits short-term from NAND shortage, but long-term valuation severely bubble-like.

  Strengths

  Very strong short-term profitability, abundant free cash flow, expected to pay huge special dividend in FY2027.

  Weaknesses

  Lack defense means against China low-price competition; Long-Term Agreements (LTA) have risk of being breached in deep bear markets.

  Comparison

  Compared to Samsung, SK Hynix, and Micron, Kioxia as an independent entity has greater valuation elasticity, but appears more fragile when facing China capacity impact.

  Risks

  NAND price falls faster than expected; Chinese manufacturer technology breakthroughs faster than expected.

Key data

• Target Price40,000 JPYApproximately 51% decrease from closing price of 81,200 JPY on June 12, 2026
• FY2026 Estimated EPS10,013 JPYSignificantly upgraded, reflecting profit surge brought by short-term NAND price skyrocketing
• FY2028 Estimated Revenue491.1 Billion JPYExpected to drop significantly 47.4% compared to 2027, reflecting price returning to normal
• Long-term Normalized Gross Margin (GM)~35%Reasonable long-term central level considered by institutions; Current stock price implies need to reach 65% to be reasonable
• YMTC Market Share ForecastThird Largest GloballyBy 2028, YMTC expected to surpass SanDisk, etc., to become top three NAND suppliers

Impact & implications

For Kioxia shareholders, this means the current stock price carnival might be the end of a cycle. Although financial report data for the coming few quarters will be very bright (EPS soaring), this is precisely a good time to sell because the market focus has not yet shifted to the price collapse in 2027-2028. For the entire NAND industry, the report suggests that returns brought by high capital expenditure will be diluted by competition from Chinese manufacturers, and industry-wide ROE will struggle to maintain the high level above consensus expectation of 50%. Investors should be wary of trades based on "permanent shortage" or "perpetual ultra-high margin" logic.

Risks

• NAND demand growth stronger than expected, leading to prolonged price and high profit duration.
• Japanese government provides favorable policy support (e.g., Capital Expenditure subsidies).
• Company cost control effect better than expected.

What to watch

• Trend of quarter-over-quarter change in NAND contract prices in Q3 2026 (watch for significant slowdown).
• Capacity ramp-up progress of new entrants like YMTC and their market share占比 globally.
• More disclosure by Kioxia on specific terms of Long-Term Agreements (LTA) (especially base price and deposit ratios).