Hitting the AI "super cycle," VGT and others see a collective performance surge

The surge in AI computing power demand has driven a super growth in the performance of related PCB companies.

Recently, VGT's performance forecast indicates that the net profit attributable to the parent company is expected to be between 4.16 billion and 4.56 billion yuan in 2025, a year-on-year increase of 260.35% to 295%.

"With the continuous expansion of global AI infrastructure and computing power demand, the company has become a core partner for many leading technology companies both domestically and internationally, thanks to its industry-leading technological capabilities, quality capabilities, delivery capabilities, and global service capabilities," VGT stated. "In key areas such as AI computing power, data centers, and high-performance computing, multiple high-end products have achieved large-scale mass production, driving the product structure to upgrade towards high value and high technical complexity, significantly increasing the proportion of high-end products and promoting rapid growth in the company's performance."

The explosion of VGT is not an isolated case; the AI computing wave is spreading throughout the entire PCB industry chain.

Dazhu CNC expects a year-on-year net profit growth of 160.64% to 193.84% in 2025, mainly driven by the demand for high multi-layer HDI boards from AI servers and high-speed switches; in 2025, WUS's net profit is expected to grow by 48% year-on-year, also benefiting from the demand for PCBs from emerging computing scenarios such as high-speed computing servers and artificial intelligence.

The industry expansion race has already begun.

WUS's high-end PCB expansion project for AI chips has commenced construction in late June 2025, with trial production expected to begin in the second half of 2026. VGT is advancing its Hong Kong IPO, and the first phase of the upgrade and renovation of its Thailand factory was completed last March, with the second phase also nearing completion.

However, Goldman Sachs believes that developing and producing the latest generation of AI servers requires PCBs that not only demand extremely high R&D investment but also face a significant capital expenditure burden, which will effectively limit the number of new entrants.

As a result, the existing leading companies are facing a relatively mild market competition environment and are expected to continue benefiting.

According to Goldman Sachs' estimation model, the global AI server PCB market size is expected to grow from approximately $3.1 billion in 2024 to $27.1 billion in 2027