top of page

How PCB Adapts to the Development of AI?

In 2024, artificial intelligence and AI will develop rapidly, with Google Xilinx, Synopsys, Many major companies such as Intel, Cadence, and Siemens are researching and developing applications of artificial intelligence.


In the fourth quarter of 2023, Microsoft's AI services contributed approximately 6 percentage points to the revenue growth of its Azure and other cloud services, a significant increase from the previous quarter's 3 percentage points, while Github Copilot had over 1.3 million paying users, a 30% increase compared to the previous quarter. In the field of cloud services, Microsoft, Google, and Amazon achieved year-on-year growth of 20%, 26%, and 13% respectively, with Microsoft and Google's growth exceeding Bloomberg's expectations, while Amazon's growth was slightly lower than expected.


pcb-for-ai

According to the composite growth rate chart of PCB sub sectors from 2018 to 2023E according to CPCA statistics, it can be seen that communication and servers have become the key tracks for rapid growth in the PCB industry in the next cycle after PCs and smartphones.


The booming development and widespread application of AI technology have led to an unprecedented demand for high-performance computing chips and AI servers.


1. The requirements are of high data volume and high transmission speed, which places higher demands on the number of layers of PCB and the material grade of CCL used for high signal transmission.


2. AI is computationally intensive, resulting in two major sources of heat generation. Therefore, PCB design needs to consider more efficient heat dissipation solutions. For example, improving PCB design layout, adding heat sinks or heat dissipation channels to achieve this.


3. Power management capabilities is improving. When a large number of GPUs or TPUs need to be deployed, more complex power management designs are required to ensure power stability.


4. The PCB board of AI servers may require higher component density and more layers due to its four board densities and levels, therefore HDI technology needs to be introduced. The introduction of HDI technology will increase the company's process flow and also enhance the added value of PCB technology.


Faced with the requirements of AI technology, from PCB design to production equipment, there are constant updates and iterations.


In the context of the incremental market of AI servers, the upstream and downstream industrial chains are undergoing rapid upgrades and iterations to meet more complex, high-performance, and high-density computing demands.

2 views0 comments

Comments


bottom of page