Recently, server concept stocks have continued to attract attention in the market, especially in the AI server sector, which is truly worth a deeper look.

First, let's discuss the industry structure. The server industry chain is roughly divided into three layers: core complete machine assembly (such as Foxconn, Quanta, Wistron), infrastructure (Vertiv, Chihong, Shuanghong, and other thermal management companies), and key components (such as Taiguang Electric, Goldsun Electric, Qincheng). Each layer is being reshuffled under the AI wave.

Speaking of which, Foxconn’s vertical integration capability is indeed unmatched, covering everything from components to complete machine assembly. NVIDIA’s Blackwell cabinet system is mainly produced by them, holding over 40% of the global AI server market share. Revenue for 2025 is projected to surpass NT$8.1 trillion, with EPS already reaching NT$10.38 in the first three quarters, setting a new record for the same period. Management says that by 2026, AI-related business will reach the trillion-dollar scale, and the growth potential is still quite large. However, the P/E ratio is already high, so caution is needed regarding risks.

Quanta adopts a different approach. Their R&D capability is recognized as the strongest in the industry, mainly serving mega clients like Google, AWS, and Meta. In Q3 2025, revenue broke NT$500 billion, an increase of over 20% year-on-year, with quarterly net profit exceeding NT$15 billion. The full-year revenue is approaching NT$1.9 trillion, with accumulated net profit of NT$68 billion. Morgan Stanley recently gave an “Outperform the Market” rating with a target price of NT$330.

Wistron, in my opinion, is the most focused. 100% of their business is in data centers, fully dedicated to ultra-large cloud service providers. They have unique advantages in high-end assembly of NVIDIA GB200/GB300 series and AMD MI series, and their liquid cooling technology integration is very deep. For 2025, their consolidated revenue reached NT$950.6 billion, a 163.68% increase year-on-year, with EPS as high as NT$275.06. The most impressive is that their order visibility extends to 2027, meaning that the growth for 2026 is basically locked in.

On the US stock side, Celestica and Vertiv are also worth noting. Celestica is a Canadian EMS giant, with differentiated advantages in manufacturing Google TPU and 800G switches. In the first three quarters of 2025, revenue was US$3.19 billion, up 28% year-on-year, with non-GAAP EPS up 52%. Wall Street’s average target price is US$374.50, with a potential upside of 22%. Vertiv, a leader in liquid cooling technology, reported third-quarter net sales of US$2.68B, up 29% year-on-year, with a backlog of orders reaching US$9.5 billion. The target price is US$206.07, with a potential upside of 27%.

However, it’s important to clarify that AI server concept stocks have already accumulated huge gains, and many leading stocks still have high P/E ratios. If signs of an AI bubble burst appear, or if investors shift from growth to profit validation, it could trigger a sharp correction. The market concentration is also quite high, so risks are significant.

Looking ahead to 2026, several variables should be monitored. First, whether cloud service providers’ investments in AI infrastructure can meet expectations. Second, progress in non-x86 architectures and self-developed ASIC chips, which could change the traditional server market landscape. Additionally, the development of edge AI and agent systems is worth watching. Changes in US regulations, tariffs, and rising electricity costs will also impact corporate gross margins. Most importantly, whether the market can shift from a speculative bubble to creating real commercial value.

In summary, server concept stocks do have opportunities, but caution is necessary when entering now. Companies with liquid cooling technology, high-density computing solutions, and deep collaborations with chip giants may continue to be favored, but valuation risks are already quite evident. It’s recommended to clearly assess your risk tolerance before making decisions.