The spillover effects of AI demand are prompting a significant revaluation across multiple sectors. During the early trading session, the fiberglass manufacturing sector experienced a collective surge. With the exception of Changhai shares, which rose over 13%, all other stocks in the sector hit the daily limit-up, a highly unusual market phenomenon. The primary drivers behind this sharp price increase appear to be a combination of price hikes and surging demand.
What triggered this movement? Analysis indicates that on February 4th, leading fiberglass producers such as Guangyuan New Materials and International Composites implemented another round of price increases for electronic cloth. This new round featured a larger price hike and a shorter cycle, signaling that the supply tightness for electronic cloth is spreading from high-end to standard products. Furthermore, the explosive demand for AI chips has caused capacity constraints for CTE cloth. Manufacturers like NVIDIA and AMD are heavily utilizing high-end fiberglass cloth, which is crowding out capacity previously allocated to consumer electronics clients.
The fiberglass manufacturing sector consists of a limited number of stocks, making a sector-wide rally with stocks hitting limit-ups or gaining over 10% a rare event. In early trading today, International Composites hit the 20% limit-up just six minutes after the market opened. Changhai shares advanced over 13%, while Honghe Technology, Zaisheng Technology, Sinoma Technology, China Jushi, Jiuding New Materials, and Shandong Fiberglass all reached their daily limit-up prices.
The revaluation of the fiberglass sector remains closely tied to both price increases and AI. A major overseas BT materials manufacturer, a downstream customer for CTE products, has explicitly stated that CTE prices were raised by 20% as of January 2026, and inventories for traditional electronic cloth are extremely low. One institution pointed out that since February, China Jushi's inventory has been under 15 days, with an absolute value slightly below 40 million meters, against an estimated annual shipment volume exceeding 1.1 billion meters for 2026.
Additional research reports highlight that specialty electronic cloth includes types such as Low-Dk cloth, quartz fiber cloth, and Low CTE fiber cloth. In the era of computing power, both AI hardware and end-user devices demand higher performance from chip materials. Low-Dk cloth is used for motherboard substrates, while Low CTE cloth is utilized for chip packaging substrates. The explosion in AI demand has led to large-scale volume increases for specialty fiberglass. Coupled with high barriers to supply, a situation of demand outstripping supply has emerged across the board.
Analysis suggests that global supply constraints are fueling optimism about pricing flexibility. Current estimates indicate the largest supply gaps exist for Low CTE electronic cloth and second-generation Low-Dk electronic cloth, with Low CTE cloth already in a phase of continuous price increases. The acceleration of domestic substitution positions AI electronic cloth as the next growth opportunity for industry leaders.
The wave of price increases is spreading beyond just "electronic cloth." According to media reports, MLCCs, often called the "rice of the electronics industry," are undergoing a new round of value reassessment and capital interest, driven by the dual forces of exploding AI computing demands and the widespread adoption of electric vehicles.
Recent spot prices for MLCCs in the South Korean market have surged by nearly 20%, indicating that this fundamental component is transitioning from traditional cyclical fluctuations to a path of structural growth guided by technological dividends. This is not merely a short-term price rebound but a fundamental shift in industry dynamics.
The application of AI technology is drastically altering the quality and quantity of demand for MLCCs. The usage in traditional servers is already significant, but powerful AI servers, due to their high power consumption and stringent voltage stability requirements, use over three times more components than standard servers, with a single device potentially requiring up to 25,000 units. High-end server racks utilizing architectures like NVIDIA's GB200 see usage counts for premium MLCCs starting in the hundreds of thousands.
Industry leader Murata Manufacturing predicts that demand from AI servers alone will grow 30% annually, with the market size reaching 3.3 times its current level by 2030. Additionally, the demand for MLCCs in new energy vehicles is up to 10 times greater than in traditional internal combustion engine vehicles. Advancements in smart driving systems and powertrain architectures are pushing automotive-grade products toward higher-end technologies like miniaturization and high-voltage resistance.
Analysis suggests that although the expansion cycle for MLCC production is relatively short and the technical barriers are lower compared to memory chips, which may limit price increases, the core future opportunity for companies that can enter the high-end AI and automotive supply chains is no longer just price volatility. Instead, the key lies in achieving qualitative transformation amidst the waves of substitution and high-end upgrades, enabling them to navigate industry cycles and share in long-term market benefits.