Summary of Key Points
The MLCC (Multilayer Ceramic Capacitor) sector in the A-share market has seen a dramatic surge recently: The sector index has nearly doubled since April, with leading companies such as Fenghua High-Tech and Sanhuan Group seeing their stock prices double since May, reaching record highs. The reason behind this is the explosive demand for AI servers, which has significantly increased both the usage and price of MLCCs. AI servers require about 10 times more MLCCs than regular servers, driving the value of a single server cabinet from several thousand dollars to over 20,000 dollars. Global giants like Murata and Samsung are overwhelmed with orders and are even raising prices to expand production. The market is now debating whether MLCCs can transform from cyclical stocks that experience price fluctuations into growth stocks with long-term growth potential. Domestic manufacturers are also seizing the opportunity to gain market share and target the high-end market.
1. How Rapidly Has the MLCC Sector Grown?
On June 2, the MLCC sector rose by over 9% in a single day, with Dalikepu hitting a 20% limit up, and Fenghua High-Tech rising by over 8%. Even more impressive figures include:
- Fenghua High-Tech’s stock price increased from 24.81 yuan on May 6 to 59.91 yuan in just 20 days (a 138% increase), reaching a record high.
- Sanhuan Group’s stock value has risen by 155% since April, with a total market capitalization exceeding 250 billion yuan.
- The Dongfang Fortune MLCC index has increased by 94% since April, nearly doubling in value.
However, Torch Electronics issued a risk warning: Its self-produced MLCCs account for only 17% of its revenue, and its computing-related businesses are still in the development stage, so the impact on earnings is minimal. Investors should not blindly chase these high prices.
2. Why Are MLCCs in Such High Demand Due to AI?
The demand for MLCCs has increased dramatically due to AI:
- Increased Usage: Regular servers use 1,800–2,500 MLCCs, while AI servers require up to 20,000 MLCCs (an 8–11 times increase). With NVIDIA’s new Rubin architecture, the number of MLCCs needed per board has further increased from 6,500 to 12,000.
- Increased Value: The value of MLCCs in a single server cabinet has risen from $3,000 in the H100 era to $12,000 for the GB200 model and $22,000 for the RubinVR200 model (more than a 7-fold increase), with potential to exceed $40,000 by 2027.
- High-End Shortage: AI servers require MLCCs that are smaller, higher in power, and more heat-resistant, leading to a severe shortage of these high-end products.
3. Global Giants Are Overwhelmed by Orders and Raising Prices to Expand Production
The actions of global MLCC leaders demonstrate the industry’s booming growth:
- Murata (Japan): AI-related orders exceed its production capacity by 200%, with an overall capacity utilization rate of 90–95%. It plans to raise product prices by 5–10% in 2026 and increase server revenue by 85–90%, while investing 80 billion yen in expanding high-end production lines.
- Samsung Electro-Mechanics (South Korea): Revenue from AI servers increased significantly in the first quarter, and the company has stated that demand for high-end products will continue to rise, with supply shortages unlikely to be resolved, leading to price increases of 5–10%.
In short, giants are struggling to meet demand and are forced to raise prices and invest in additional production capacity.
4. Can MLCCs Repeat the “Growth Myth” of Storage Chips?
The success of storage chips, which shifted from cyclical stocks to growth stocks due to sustained demand, could be replicated for MLCCs:
- Similar Logic: High-end AI MLCCs require a larger share of production capacity (e.g., several dozen times more than general-purpose MLCCs). Overseas giants prioritize high-end production, leaving a shortage in the mid-to-low-end market. This is not a traditional inventory replenishment cycle but a structural gap.
- Future Trend: Industry analysts predict that MLCCs will see a pattern of high-end shortages and stable demand for mid-to-low-end products. Prices for high-end AI and automotive-grade products are expected to rise, while mid-to-low-end prices will remain relatively stable. If AI demand continues to grow, MLCCs could break away from their cyclical nature and become growth stocks.
5. Opportunities for Domestic Manufacturers
The focus on high-end AI products by overseas giants creates opportunities for domestic manufacturers:
1. Gaining Market Share in Mid-to-Low-End Markets: As overseas manufacturers shift production to high-end products, domestic manufacturers can expand their market share in the mid-to-low-end segment.
2. Targeting High-End Products: Rising industry prices have improved domestic manufacturers’ profitability, allowing them to invest in research and development for high-end MLCCs used in AI servers and automotive applications. For example, Fenghua High-Tech is expanding production of high-end resistors and inductors, and Guoci Materials is increasing production of MLCC powder for AI servers.
If domestic manufacturers can overcome technical challenges in high-end products, their performance could significantly improve in the future.
Conclusion
MLCCs are experiencing rapid growth due to AI demand, but there are also risks to consider. For instance, some companies have a low proportion of MLCC sales in their revenue, and it remains uncertain whether the industry can truly escape its cyclical nature. However, with long-term demand for high-end MLCCs driven by AI, domestic manufacturers could see significant opportunities. Investors should remain rational and not be solely focused on short-term price gains.