Large Gains in Fiberglass Stocks: Shortage of Electronic Fabric Brings New Constraints to AI Computing Power!

Large Gains in Fiberglass Stocks: Shortage of Electronic Fabric Brings New Constraints to AI Computing Power!

```

A-share glass fiber sector surged collectively. On February 11, International Composites rapidly hit a 20cm daily limit, with Honghe Technology, Sinoma Technology, Shandong Fiberglass and other stocks also hitting their limit. Looking over a longer period, the Wind Glass Fiber Index has gained more than 74% cumulatively since December 1, 2025.

Electronic cloth refers to electronic-grade glass fiber cloth used in the electronics industry, and is a key fundamental material for printed circuit boards (PCBs). High-end electronic cloth is an indispensable core skeleton material for high-performance chip substrates and PCBs. Driven by the explosive growth in AI chip demand, the high-end electronic cloth market is facing increasingly significant supply shortages.

The direct cause of this market surge comes from another recent price hike in electronic cloth. The supply of high-end electronic glass fiber cloth produced by Nitto Boseki is increasingly tight, and the market is even seeing a "one cloth hard to find" situation. To ensure stable supply of core raw materials, Nvidia CEO Jensen Huang recently visited Japan in person to negotiate cooperation with Nitto Boseki, and Apple is also actively communicating with relevant Japanese government departments to seek supply chain support.

On February 4, industry leaders International Composites and Guangyuan New Materials raised prices for electronic cloth, with regular electronic cloth prices rising more than 10%, breaking the 10,000 yuan threshold. This is the fourth price increase since early 2025, and it is the largest and fastest, reflecting the overall supply/demand tension for electronic yarn/cloth. The mainstream price quotes for regular electronic yarn have reached 10,300-10,700 yuan/ton, and mainstream prices for 7628 electronic cloth have risen to 5.1-5.5 yuan/meter.

According to Huatai Securities estimates, this price hike is expected to bring a net profit increment of about 0.5 yuan/meter, contributing incremental net profit of about 440 million yuan, 110 million yuan, and 110 million yuan respectively to China Jushi, International Composites, and Honghe Technology, accounting for about 13%, 34%, and 50% of their 2025 net profit attributable to shareholders.

Behind this round of price increases are both supply constraints and demand growth. Driven by the construction of AI servers, high-speed network equipment, and other computing power infrastructure, demand for electronic cloth continues to climb; while on the supply side, high investment costs and technical barriers mean limited capacity expansion. The industry has entered a new business cycle. Tight electronic cloth supply may create new supply chain constraints on computing infrastructure development.

Limited Supply Increase, Rising Costs Suppress Expansion

Based on industry forecasts, the net increase in regular electronic yarn supply by 2026 may be less than 10%. According to Zhuochuang Information statistics, during 2023 to the first half of 2025, apart from a few capacity replacements in regular electronic yarn, no new production lines have come online. New capacity for 2026 will mainly come from China Jushi's Huai'an Project (Phase I of 50,000 tons expected in Q1, Phase II of 50,000 tons planned for the second half of the year) and Kingboard Clearwater Line 7 (70,000 tons expected in the second half). However, considering International Composites will have two kilns entering cold repair during the same period, the actual net increase in capacity as a proportion of current operating capacity is expected to be less than 10%.

On the supply side, there are also dual constraints of structural shift and rising costs. Driven by tight supply and high profitability of high-end electronic cloth, some regular electronic yarn production lines have shifted to high-end products, further compressing regular yarn supply. At the same time, since 2026, upstream precious metal prices have risen significantly. As a key leak plate material, the price surge in platinum-rhodium alloy has pushed up both kiln investment and daily operating costs. According to Huatai Securities, precious metal leak plates account for close to 40% of total investment in electronic yarn production lines, and the cost pressure may further suppress industry capacity expansion.

Capacity expansion is also limited by extremely high capital and technology thresholds. Basic investment for a single electronic yarn kiln exceeds 500 million yuan, while investment in high-end production lines even goes over 1.5 billion yuan, with a construction-to-production cycle of over two years, and single standard line equipment investment often exceeds 500 million yuan. Such large-scale, long-cycle, and irreversible heavy asset investment forms substantial barriers for new industry entrants.

AI Computing Power Upgrades Drive Demand Growth

Market demand is experiencing a structural rebound. As a core base material in the copper clad laminate (CCL) and printed circuit board (PCB) industry chain, terminal demand for electronic yarn/cloth is being steadily supported by orderly recovery in traditional sectors such as home appliances, new energy vehicles, and consumer electronics, driven by "old-for-new" and "new national subsidies" policies. Demand for regular electronic cloth is expected to rise steadily as a result.

An even more significant driving force comes from high-end computing hardware upgrades, represented by AI servers. As AI servers evolve from traditional CPU architecture to GPU cluster architecture, the number of PCB layers commonly increases to over 20, demanding higher performance in base materials.

For example, Nvidia’s GB300 servers have PCB layer counts of more than 16, with each server requiring about 18 to 24 meters of quartz electronic cloth (Q cloth), approximately five times higher than usage in traditional servers. Next-generation high-speed optical modules (1.6T/3.2T) and switches (224G ports) also rely on high-performance electronic cloth to ensure transmission integrity and stability.

In terms of material performance, the dielectric properties of electronic cloth directly affect signal transmission efficiency. Every 10% decrease in dielectric constant can double transmission speeds. Third-generation quartz electronic cloth (Q cloth) has lowered dielectric constant to a range of 2.2–2.3, controlled dielectric loss factor at 0.001–0.003, and thermal resistance over 600°C. Such materials can significantly reduce signal attenuation and distortion in high-frequency transmission, providing faster and more stable signal transmission for AI computing equipment.

Accelerated Domestic Breakthroughs in High-End Market

In the high-end electronic cloth market, Japanese companies still dominate. Nitto Boseki, Asahi Kasei, and Asahi Glass together control nearly 70% of the global high-end electronic cloth market. Nitto, with long-term accumulation in NE-glass and T-glass core glass material formulations, has been investing in R&D since the 1990s, building up over 30 years of technology, giving it a first-mover advantage in the current industry boom.

According to media citing Professor Zeng Huidan of East China University of Science and Technology, Nitto has a significant advantage in the second-generation low dielectric (LDK/LCTE) electronic cloth sector, achieving stable mass production for forty years, with yield rates above 95%. Domestic companies have achieved technological breakthroughs in recent years, but mass production yield rates generally range from 70% to 80%, and production capacity is less than 20% of Nitto's. She points out, China already ranks among the global leaders in mid-to-low end electronic cloth, but in the highest-end products, it is still transitioning from technology breakthroughs to large-scale mass production.

Domestic substitution is accelerating. In 2021, Honghe Technology achieved volume production of 9-micron ultra-thin electronic cloth, breaking foreign monopoly; the same year, Henan Guangyuan completed volume production of low dielectric cloth, becoming the first domestic firm to break through in this segment. In 2024, Taishan Glass Fiber achieved volume production of second-generation low dielectric cloth. In the more technologically demanding quartz electronic cloth (Q cloth) sector, Feilihua completed eight years of R&D and successfully launched M9 grade Q cloth in 2025 and obtained official Nvidia certification.

Tianfeng Securities points out that Q cloth industry barriers are mainly in three aspects: First, complex drawing process—quartz fibers are brittle and kiln temperatures need to exceed 2000°C; second, core equipment such as looms still rely on imports, with long lead times; third, strict customer certification, usually taking 2-3 years. High barriers also mean high added value: first-gen cloth is about 30 yuan/meter, second-gen about 120 yuan/meter, whereas Q cloth reaches 200–400 yuan/meter, with gross margins exceeding 60%.

From an inventory perspective, as of September 2025, inventories of China Jushi, International Composites, and Honghe Technology were 3.72 billion yuan, 1.92 billion yuan, and 170 million yuan respectively, equal to 82%, 89% and 77% of peak levels since 2021. The industry's overall inventory level is healthy, providing room for price transmission.

Looking to the future, Professor Zeng Huidan believes that a breakthrough in high-end capacity will face challenges before 2027, but the long-term trend of domestic substitution is clear. China has already achieved breakthroughs in key technologies, and as mass production capability and customer cooperation deepen, it is expected that in the next five to six years Chinese companies will achieve key breakthroughs in some high-end market segments and establish substantial market shares.

Risk Warning and DisclaimerThe market has risks, investments should be made cautiously. This article does not constitute personal investment advice, nor does it take into account the particular investment objectives, financial situations, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article suit their specific circumstances. Investment based on this article is at your own risk. ```