Copper interconnect technology has become a key factor in improving data center performance.
According to the research report released by Orient Securities, as the core carrier of computing power, data centers are expected to rapidly expand in the server field and the upstream and downstream industries in the next few years driven by the high demand for computing power. Based on this trend, copper connectivity solutions are expected to increase their market share in short-range connections within data centers due to their low cost and low power consumption, thereby promoting industry development. With the release of new products from Nvidia and the continuous development of domestic and foreign data center integration solutions, the copper interconnect market is expected to see significant incremental demand in the future. In addition, domestic manufacturers have the capacity for high-speed copper cables needed for AI computing power and mass production of upstream components.
Orient Securities' main points are as follows:
Artificial intelligence promotes high-speed copper cable industry demand
With the gradual maturity of generative artificial intelligence (AIGC) technology, emerging applications based on the "big data + large computing capacity" model continue to emerge, significantly driving the continuous increase in computing power demand. According to the data from the China Academy of Information and Communications Technology, it is expected that the global annual average computing power scale will grow at a rate of over 50% from 2023 onwards for the next five years. As the core carrier of computing power, data centers are expected to rapidly expand in the server field and the upstream and downstream industries in the next few years driven by the high demand for computing power. Based on this trend, copper connectivity solutions are expected to increase their market share in short-range connections within data centers due to their low cost and low power consumption, thereby promoting industry development.
Copper interconnect technology has become a key factor in improving data center performance
With the development of large-scale data centers and edge computing, the demand for copper continues to grow. Although optical fiber is widely used for long-distance communication, copper cables occupy an important position in short-distance transmission due to their high cost-effectiveness, low power consumption, and high reliability, especially in the construction of high-performance AI computing clusters, where the low cost and low energy consumption advantages of DAC copper cables are particularly significant. The DAC solution not only achieves a data transfer rate of up to 100Gbps but also alleviates the limitations of short-distance transmission. In addition, the high mean time between failures (MTBF) of DAC copper cables ensures the continuity of critical tasks in data centers. According to LightCounting's forecast, the high-speed cable market will double in the next five years, and DAC sales will rapidly increase at an average annual compound growth rate of 25%.
The NVL72 server opens up growth space for the copper connectivity market.
As a leader in the global accelerated computing field, Nvidia is considered a benchmark for the AI industry. In 2024, Nvidia released the highly anticipated Blackwell architecture, and the new products based on this architecture have attracted widespread attention in the market. Among them, the NVL72 server, as a new AI server, adopts a copper interconnect solution, including 5000 copper cables with a total length of over 3219 meters. The introduction of new technology makes the NVL72 server a model of high-performance, low-power consumption equipment. With the release of Nvidia's new products and the continuous development of domestic and foreign data center integration solutions, significant incremental demand is expected in the copper interconnect market in the future.
Domestic copper connector manufacturers are on the rise, and the future market prospects are promising.
Domestic manufacturers have the mass production capability of high-speed copper cables and their upstream components required for AI computing power: Woer Heat-shrinkable Material is a leading enterprise in high-speed communication cables, and its subsidiary Leting Zhilian is deeply involved in the wire industry with strong R&D and production capabilities. It has developed samples of 224G high-speed communication cables, and the sampling of 1.6T high-speed communication cables is in progress; Dongguan Dingtong Technology focuses on communication connectors and is an important supplier of Amphenol and Molex. Its 112G products are in mass production, and 224G products are also being trial-produced in small batches. The utilization rate of its Dongguan factory is relatively high, and the expansion of 56G and above high-speed backplane connectors is in progress; Huafeng Technology has mastered the core technology of connectors and successfully developed 112G and 224G high-speed backplane connectors. High-speed cable modules have begun mass production and delivery; Jingda shares focus on special electromagnetic wires for more than 30 years, and the company's silver-plated wires serve as the transmission carrier for high-frequency and high-speed signals. Customers include Amphenol, Leting, Molex, Luxshare Precision Industry, Zhejiang Zhaolong Interconnect Technology, Xinya Electronic, etc.
Recommended targets to watch: Wire manufacturers: Woer Heat-shrinkable Material (002130.SZ), Jingda shares (600577.SH), Xinya Electronic (605277.SH), Shenyu Communication Technology (300563.SZ); Connector and component manufacturers: Dongguan Dingtong Technology (688668.SH), Huafeng Technology (688629.SH), Yihua Connector (002897.SZ), Shenzhen Deren Electronic (002055.SZ); High-speed cable manufacturers: Luxshare Precision Industry (002475.SZ), Zhejiang Zhaolong Interconnect Technology (300913.SZ), FIT HON TENG (06088); Copper foil manufacturers: Guangzhou Fangbang Electronics (688020.SH), etc.
Risk warning: Technology development and implementation may not meet expectations; downstream market demand may not meet expectations; the risk of rising raw material costs; increased market competition risk; supply chain stability risk; geopolitical risk.