The growth of the server market drives the demand for PCBs. PCB is an important component of servers and a key material for the operation of servers. The substantial increase in server shipments has also expanded the server PCB market, making it the downstream market segment with the fastest compound growth rate in the PCB market. PCB is an important component of the server and a key material for the operation of the server. The substantial increase in server shipments has also expanded the server PCB market, making it the downstream market segment with the fastest compound growth rate in the PCB market. With the increasing demand for AI computing power, the PCB market is expected to usher in a recovery.

Statistics from Prismark show that the current global PCB market is growing steadily. Its output value is expected to reach US$82.1 billion in 2022, +1.5% year-on-year. Looking forward to the future, with the weakening of the macro-influence margin and the steady recovery of overall demand, the superimposition of servers and data centers, automotive electronics, AIoT (smart earphones, smart watches, AR/VR, etc.) emerging applications and technology upgrades, PCB output value is expected to grow steadily, Prismark It is estimated that the global output value will reach 101.6 billion US dollars in 2026, and the CAGR will exceed 5% in 2022-2026. Benefiting from the continuous iterative upgrade of downstream application technical specifications, the requirements for circuit boards have also continued to increase, and the output value of PCB downstream mid-to-high-end products such as HDI, packaging substrates, etc. has increased significantly. Since the birth of ChatGPT last year and the rapid increase in the number of monthly active users, AI has not only attracted the attention of global investors, but has also become a hot topic on global social networks. Compared with the high-tech distance that AI gave people in the past, this innovation made everyone realize the possibility of AI entering the lives of ordinary people. Nvidia’s financial report has exceeded expectations, and the acceleration trend of AI has been established from the source of computing power; Marvell, a major American chip design company, also emphasized that the revenue of AI-related products will more than double every year in the next two years.

In the past 10 years, the main carrier of computing power in the AI field has been mainly GPU equipment provided by foreign chip manufacturers, which are widely used in AI-related cloud products. The end-side embedded AI computing power carrier has developed from CPU, GPU, DSP to ASIC architecture, which has promoted the development of deep learning-based speech recognition, face recognition, graphic recognition, AIGC, object detection, super-resolution, ADAS and other technologies. widely used.