AI drives the rapid development of PCB industry
The development of artificial intelligence will reshape the electronic semiconductor infrastructure. The collection, calculation, training and transmission requirements of massive data will bring about the iterative upgrade of computing power and network. At the same time, AI has greatly boosted the demand for high computing power, which is expected to bring a new round of opportunities for upgrading electronic information infrastructure.
AI drives the development of PCB industry
PCB is known as the “mother of electronic products” and is used in almost all electronic products. PCB is an important part of the electronics industry to realize the electrical connection between the components in the circuit. There are a wide range of downstream applications, among which communications, computers, and automotive electronics account for over 70% of total PCB consumption. Order visibility for PCB and IC substrate production equipment has been extended to 2024.
Behind the artificial intelligence model is a huge computing power support. AI computing training and edge-side reasoning not only place high computing power requirements on GPUs, AI chips, and storage, but also require comprehensive server infrastructure support. It will drive the construction of related IDCs such as switches and servers, and the demand for graphics cards, and related PCB and interface companies are expected to continue to benefit.
According to industry research reports, in the long run, the demand for the communication sector in the PCB market is stable, and the demand for data centers and automobiles is still high. It is estimated that the compound annual growth rate of the PCB market will be 4.8% from 2023 to 2026, and the output value of the global PCB market will reach 101.5 billion US dollars by 2026.
The accelerated development of the AI industry brings long-term development opportunities for the electronics industry
According to data from Precedence Research, the global artificial intelligence chip market size will be US$16.86 billion in 2022. From 2022 to 2032, the CAGR is expected to reach 29.7%. We are optimistic about the new development opportunities brought by the accelerated development of artificial intelligence to the electronics industry, and the sub-sectors are expected to continue to benefit:
① Computing chip
AI drives up the volume and price of CPU/GPU/FPGA/ASIC, and there is huge room for domestic substitution.
It is optimistic that the storage sector will stop falling and rebound in 2023, and DDR5 memory is expected to increase in volume. According to industry estimates, the CAGR of the memory interface and supporting chip market will reach 28% from 2021 to 2028.
On the basis of limited local area network communication, the server side has a lot to do. Industry organizations predict that the market size of Ethernet switching chips for data centers in China will reach 12.04 billion RMB in 2025.
④Power management chip
Demand is expected to pick up in the second half of 2023. With the growing demand for cloud computing, artificial intelligence, and high-power processors and accelerator cards, the accelerated server voltage has shifted to 48V, and DC-DC has benefited significantly.
AI catalyzes the demand for computing power, and the volume and price of server motherboards and carrier boards are rising. The demand for AI is rising, and the tide of server PCB/CCL upgrade is coming. Computing power has greatly increased the requirements for CPU/GPU, and advanced packaging highlights the value of carrier boards.
⑥ Optical chip
With the explosive growth of global data volume, optical communication is gradually rising. According to Lightcounting data, the global optical module market will exceed 20 billion in 2027, and optical chips are the core components of optical modules.
The rapid iteration of AIGC services such as ChatGPT brings a new round of communication infrastructure upgrade
On the one hand, ChatGPT has greatly boosted the demand for high computing power, driving server growth and increasing the proportion of AI servers. According to TrendForce data, as of 2022, the annual shipments of AI servers equipped with GPGPU will account for nearly 1% of the total servers. In 2023, it is estimated that with the support of ChatGPT, the annual growth rate of shipments will reach 8%, and the CAGR will reach 10.8% in 2022-2026.
On the other hand, ChatGPT drives the upgrading of the cloud computing industry chain, accelerates the commercialization of 800G optical modules, and drives the demand for advanced packaging technologies such as CPO. According to LightCounting, the global top 5 cloud manufacturers spent US$1.4 billion on Ethernet optical modules in 2020, and their expenditures are expected to exceed US$3 billion by 2026. 800G optical modules will dominate this market segment from the end of 2025.
The characteristics of the ChatGPT large model have brought about a sharp increase in computing power requirements. The GPT3 model alone requires thousands of A100 chips for more than a month of training. As the carrier of computing power and transmission hardware, servers and switches will greatly increase the consumption of components such as PCBs and connectors.
At the same time, with the explosion of digital economy, 6G, artificial intelligence and other industries this year, PCB is considered by the industry to usher in a boom in demand.