Key facts
- Asian AI stocks in Japan, South Korea, and Taiwan are experiencing heightened volatility.
- Retail investors are contributing to this volatility through increased leverage.
- Concentration in top tech names is driving earnings growth, but also poses a risk.
- Demand for AI infrastructure, particularly memory chips and advanced manufacturing, exceeds current capacity.
- Industry forecasts suggest memory chip shortages will continue for several years.
- Supporting companies in the AI supply chain are also demonstrating significant growth.
Asian equity markets, particularly in Japan, South Korea, and Taiwan, are experiencing increased volatility as retail investors amplify the rally in AI-related stocks through higher leverage. This concentration in a few key names, while driving significant earnings growth, raises concerns about the sustainability of the current upward trend.
The strong performance of Asian tech stocks is largely attributed to the concentrated earnings growth from the top three technology companies, which are expected to contribute substantially to the overall earnings surge for the region. This has led to a debate between bulls, who see it as proof of AI demand translating into real earnings, and bears, who view it as a crowded rally with limited further upside for the biggest winners.
Despite these concerns, earnings beats continue to outpace misses, and management commentary consistently points to demand exceeding supply, particularly in the manufacturing and infrastructure layers of the AI ecosystem. Taiwan Semiconductor Manufacturing Company (TSMC) serves as a key indicator, with strong demand for its advanced 3-nanometer technology and packaging services outstripping its current capacity. The company anticipates the global semiconductor market to surpass $1.5 trillion by 2030, with AI and high-performance computing accounting for a significant portion.
The demand for AI infrastructure is projected to grow exponentially, with AI accelerator wafer demand expected to increase elevenfold from 2022 to 2026. This, coupled with Big Tech's substantial planned spending on AI rollouts, underscores the ongoing need for manufacturing and packaging capacity. Consequently, investors are advised to favor companies involved in advanced nodes, packaging, and the physical supply chain that cannot be rapidly scaled.
The rally's breadth is also expanding beyond the headline names, with supporting suppliers in areas like optical interconnects and printed circuit boards (PCBs) posting notable revenue and earnings growth. Companies such as Zhongji Innolight, Suzhou TFC Optical Communication, and Unimicron Technology are demonstrating growth figures consistent with robust AI demand. This broadening participation suggests that the bottlenecks are spreading outward from the core chip to essential components that scale with cluster size, indicating a more widespread impact of the AI buildout.