On May 20, computer manufacturer Dell announced a new series of personal computers optimized for AI use. During a speech in Las Vegas, the CEO mentioned, "We will deliver in large quantities this year, and by next year, it will become standard equipment."

 

In the past, it was not uncommon for PC users to connect to cloud servers via the internet to perform AI-related computations. Broadly speaking, it is not a new experience for users to enjoy the benefits of AI on their PCs. However, the recent hot topic of “AI PC” specifically refers to Edge AI, meaning that most of the AI computations are carried out locally on the PC itself.

According to Intel and Microsoft, an AI PC must have the following features: 1) the ability to run Microsoft Copilot, 2) a dedicated NPU (Neural Processing Unit), and 3) a dedicated Copilot button. Although this definition is not absolute, it is significant enough in the current Wintel-dominated PC ecosystem to influence the majority of PC brands. In other words, most future PC models are expected to follow this definition.

However, concerning PCB, the current definition emphasizes that the NPU is the key influencing factor. According to IDC's classification, AI PC are categorized based on TOPS computational power as follows: Hardware-enabled (<40 TOPS), Next-generation (40-60 TOPS), and Advanced (>60 TOPS). Currently, mainstream chips from Qualcomm, Apple, AMD, and Intel belong to the first category. It is estimated that they will continue to account for over half of the market share until 2026. In other words, the mainstream system specifications that PCB products will face in the next two to three years will be designed based on this standard.