摩根大通：日本MLCC股进入“利润验证期”，10-12月成关键窗口

TL;DR

• Feedback from JPMorgan's roadshows in Hong Kong and Singapore in late June indicates that investor interest in Japanese MLCC and passive component stocks remains high, but valuations have clearly diverged.
• Bullish investors still favor Murata, but many funds are now managing risk by going long on Murata while shorting Taiyo Yuden or TDK.
• In the short term, it is becoming increasingly difficult to push stock prices higher on AI server demand alone. The next key catalysts will be earnings reports, demand in October-December, and 2027 price negotiations.

Feedback from JPMorgan's Asian roadshow for the electronic components industry, published on June 30, shows that after the significant rally in Japanese MLCC and passive component stocks driven by AI server demand, institutional investors' views on sector valuations are becoming increasingly divided.

From June 22 to 26, JPMorgan met with investors in Hong Kong and Singapore. Most meeting questions focused on passive component manufacturers, particularly MLCC-related companies, including Murata Manufacturing, TDK, Taiyo Yuden, Nichicon, and Nippon Chemi-Con. Investors also focused on other AI server chain beneficiaries such as Ibiden, Rohm, MinebeaMitsumi, Alps Alpine, and Hirose Electric. The report also noted a significant increase in hedge fund inquiries regarding potential short targets.

The core change in this feedback is that the market is no longer just discussing "how much demand AI servers will generate," but is starting to ask: Is this demand sufficient to support the already significantly inflated stock prices?

Stock Prices Have Run Ahead, Valuation Divergence is Widening

Since April, the valuations of many Japanese electronic component stocks have been significantly above their historical ranges. The report states that investors are currently divided into two main camps: one believes AI server demand is creating a new normal for growth, while the other believes current stock prices represent excessive pricing.

Passive component companies, including MLCC, aluminum electrolytic capacitors, and crystal devices, have been the most closely watched sector in this trading cycle. As AI server power consumption increases and board design complexity rises, demand for high-capacity, high-reliability MLCCs has increased, leading to rapidly rising growth expectations for related manufacturers.

However, this rapid price increase has also blurred consensus. The report points out that some investors believe that after the short-term price surge, it is difficult for the market to determine where the true consensus expectations lie. In other words, while strong demand has become consensus, how much future growth is already priced into the stocks is becoming a new point of divergence.

In terms of stock performance, from late March to mid-June, Murata rose approximately 35%, Taiyo Yuden rose approximately 50%, while the TOPIX index rose about 17%. Passive component stocks have significantly outperformed the broader market, becoming one of the more crowded assets in the Japanese AI trade.

Specifically, while stocks like Taiyo Yuden, Nichicon, and Nippon Chemi-Con have risen notably, their current quarterly profit levels remain low. Consequently, many investors worry that if April-June results fail to demonstrate simultaneous improvements in orders, prices, and profit margins, there could be downward pressure on stock prices after the earnings announcements.

Murata Remains the Core Position, Taiyo Yuden and TDK Become Hedging Tools

In terms of specific positions, Murata remains the most favored long. This is due to its stronger position in high-end MLCCs, where its customer base and product specifications better align with the logic of AI server demand upgrades, making it more easily viewed as a core beneficiary by long-term capital.

However, with stock prices overall at high levels, investors are also using relative value trades to manage risk. The report shows that some early investors have begun taking profits on Taiyo Yuden while shifting to a pairs trade of long Murata and short Taiyo Yuden. Other investors are holding Murata and waiting for Taiyo Yuden to pull back before re-entering.

This indicates that capital has not completely exited the MLCC sector but has shifted from "buying the entire sector" to "buying the leader and selling the high-beta names." Taiyo Yuden is viewed as one of the direct beneficiaries of AI server MLCC demand, and its stock price has shown higher elasticity earlier; however, greater elasticity also means higher pressure to deliver on earnings. If order growth, price increases, or capacity utilization improvements do not quickly translate into the profit and loss statement, short-term capital is more likely to choose to lock in gains.

TDK's position is more nuanced. The report notes that some investors view TDK as a latecomer in AI server MLCCs; however, the market also believes that the supply-demand balance for power inductors related to AI server power, especially thin-film metal inductors, is tightening, potentially leading to price increases. Furthermore, TDK has potential upside from its increasing market share in LFP backup battery units (LFP BBU) and HDD heads.

Therefore, TDK could be seen both as a spillover beneficiary of the AI server chain and as a potential hedging tool in pairs trades after its valuation has risen. The report mentions that many investors are employing a combination of long Murata and short TDK, indicating that the market has begun repricing across different AI component beneficiaries.

Similar divergence is appearing in other electronic component stocks. Ibiden remains a beneficiary of AI server substrate demand, but its stock price already reflects earnings expectations for FY2030 or even further out. Interest in Rohm is rising, driven by the increased value of its Kioxia stake and shortages in 80-100V Si-MOSFETs.

MinebeaMitsumi benefits from demand for server fan bearings and BBU protection modules, but its valuation is no longer cheap. Alps Alpine has become a target for some hedge funds seeking short ideas.

Next Steps: Earnings, October-December Demand, and 2027 Price Negotiations

JPMorgan is relatively restrained on short-term stock performance. The report believes that while MLCC-related stocks have risen significantly since April, analyst earnings forecast revisions over the past three months have been limited. Therefore, during the July-September period, it is unlikely that AI server MLCC expectations alone will drive stock prices significantly higher; instead, these stocks are more likely to enter a range-bound consolidation phase.

The real next catalyst may appear in October-December.

The report suggests that demand for high-end AI server MLCCs may become more apparent in October-December. As 2027 price negotiations progress, further tightness in the MLCC supply chain would make price increase expectations more concrete, potentially reigniting stock price momentum.

Beyond MLCC, Nichicon and Nippon Chemi-Con are also being re-evaluated within the AI server power supply chain. The market focus is on whether their price increases merely pass through rising aluminum costs or if they can achieve more substantial price increases for higher-specification products. As server power architectures evolve towards 400V and 800V HV-DC, demand for long-life, high-reliability aluminum electrolytic capacitors and high-voltage related components may increase.

This is the key area the market needs to validate most: AI server demand has been thoroughly discussed, but whether price increases can be realized, their magnitude, and their sustainability will directly determine whether this current cycle continues or enters a phase of valuation digestion.

Hedge funds beginning to search for short ideas does not necessarily indicate an immediate peak for Japanese passive component stocks. More accurately, the market is transitioning from a one-sided bet on AI demand into a more detailed phase of valuation and profit validation. Investors will continue to buy leaders with higher certainty while also seeking targets perceived as overvalued, with insufficient profit delivery, or excessively theme-driven for hedging purposes.

For Murata, the key is whether high-end MLCC demand can continue to support its leadership premium. For Taiyo Yuden, the key is whether its profit and loss statement can catch up with its stock price elasticity. For TDK, the key is whether power inductors, LFP BBU, and HDD heads can provide new growth support.

This roadshow feedback reflects the real-time sentiment of some institutions and hedge funds. It is not a formal earnings forecast nor does it represent the consensus of the entire market. However, it clearly reveals a shift: AI server demand has propelled Japanese passive component stocks to high levels. The market's next focus will no longer be solely on the strength of the demand narrative, but on whether earnings reports, orders, and price negotiations can justify the already-inflated stock prices.