Summary of Key Points
In the first half of the year, there was a surge in demand for AI computing power, which drove up both the volume and price of electronic fabric (a type of glass fiber fabric used for making circuit boards), leading to significant increases in the performance of leading companies in the industry. However, there is a clear differentiation within the industry: ordinary high-dielectric E-fabric (suitable for traditional electronic devices) represents a short-term cyclical benefit with low entry barriers and limited profit elasticity, and its profitability is susceptible to changes in production capacity; low-dielectric special fabric (suitable for AI computing equipment) is a long-term high-value segment with high profits and significant potential for domestic substitution, but it faces challenges in terms of technology, equipment, and certification. Overall, the industry's prosperity is expected to continue until the end of the year, but the opportunities are not equal—ordinary fabric may offer quick profits, while low-dielectric fabric is the key to determining a company's long-term success, and domestic substitution requires gradual breakthroughs.
Detailed Analysis
1. AI Computing Power Drives the Rise of Electronic Fabric
Electronic fabric is a core material for printed circuit boards (PCBs). The surge in demand for PCBs from AI servers and computing devices has directly increased the usage of electronic fabric. Coupled with low industry inventory and a shortage of high-end weaving machines globally, the release of new production capacity has been slow, resulting in a severe mismatch between supply and demand. This has led to substantial price increases, and companies have seen a boost in orders. Institutions predict that this prosperity will continue until the end of the year as capacity bottlenecks are difficult to resolve in the short term.
2. Ordinary High-Dielectric E-Fabric: Short-Term Benefits, but Vulnerable
Ordinary E-fabric (such as 7628 and 2116 specifications) is commonly used in mature devices like smartphones and traditional servers. China has substantial production capacity and mature technology, and almost all manufacturers can produce it. The price increase is not due to a sudden surge in traditional demand but rather because leading companies have shifted their weaving machine resources to more profitable low-dielectric fabric, reducing the supply of ordinary fabric. However, these benefits are fragile:
- The supply shortage is due to internal reallocations, not a sudden increase in demand; if leading companies return to producing ordinary fabric, supply will quickly recover.
- The entry barriers are low, and any company with the funds to purchase weaving machines can produce it. Once additional capacity is added, prices may fall.
- Small and medium-sized manufacturers, lacking their own upstream yarn supply, will see their profits eroded by higher raw material costs and are at the mercy of market trends.
3. Low-Dielectric Special Fabric: A Critical Need for AI, High Profits, but High Barriers
AI equipment requires fast signal transmission, and ordinary fabric has high signal loss and latency issues. Low-dielectric fabric solves these problems, with much higher unit prices and gross margins (for example, while ordinary fabric sells for a few yuan per meter, low-dielectric fabric can sell for dozens of yuan per meter). The increased demand from AI is largely focused on this type of fabric. However, entering the low-dielectric fabric market is challenging:
- Equipment Constraints: High-quality Japanese weaving machines are needed, with limited global production and long delivery times, and domestic machines are not yet mature.
- Mass Production Challenges: It is easy to produce samples in the laboratory, but maintaining consistent signal loss and fabric flatness during mass production is difficult, and improving yield rates takes a long time.
- Certification Requirements: AI servers have strict requirements for material reliability, and downstream customers (such as Huawei and Inspur) undergo lengthy testing processes before switching suppliers.
4. Opportunities for Domestic Substitution, but a Long Road Ahead
Previously, the low-dielectric fabric market was dominated by Japanese (Toray) and Taiwanese (Fuqiao) companies. With the surge in AI demand, overseas production capacity cannot keep up, and downstream customers are beginning to test domestic products, providing opportunities for domestic substitution. Leading Chinese companies (such as China Jushi and Honghe Technology) can produce samples and have started small-scale production, but there are still hurdles to overcome:
- Equipment: Imported weaving machines are hard to obtain, and domestic machines are still being validated.
- Mass Production: Yield rates are unstable, and costs are high.
- Certification: The certification process is lengthy, and it may take 1-2 years to secure large orders.
Therefore, domestic substitution is a gradual process, and a complete replacement is unlikely in the short term.
5. Two Different Paths, Two Different Futures
Ordinary E-fabric represents a cyclical business, with large fluctuations in performance due to short-term price increases and intense competition. Low-dielectric fabric, on the other hand, represents a growth opportunity with stable profits and significant potential. For companies, focusing on ordinary fabric can lead to quick profits in good markets but back to price competition in bad times. For long-term development, investing in low-dielectric fabric is necessary, which requires significant investment in research and development, equipment acquisition, and certification—not all companies can afford this.
In summary, the electronic fabric industry is booming, but ordinary fabric offers short-term gains, while low-dielectric fabric is the key to long-term success. There are opportunities for domestic substitution, but it requires a step-by-step approach and patience.