The business cycle of semiconductor investment

In the capital market, semiconductor stocks have long been regarded as a highly cyclical sector. Valuations expand rapidly during upturns and are often sharply corrected during downturns. However, understanding the semiconductor industry solely through short-term supply-demand dynamics or business cycles can easily lead to overlooking its deeper technological evolution. In fact, semiconductors are not a single industry but rather a collection of industries that continuously reorganize with technological specialization, application scenarios, and capital structures. To understand their long-term valuation, one must look beyond quarterly earnings and seek clues from changes in market capitalization structures.

Looking back at the structure of the top 50 global semiconductor companies by market capitalization over the past two decades, it is clear to see a shift in valuation focus. In earlier lists, integrated device manufacturers and memory companies accounted for the largest share, with market values heavily concentrated on process scale and capital expenditure capabilities. The core competitiveness at that time was who could expand production faster and reduce unit costs more effectively, leading to valuations that fluctuated dramatically with economic conditions.

Take the U.S. market as an example: NVIDIA has consistently ranked at the top of the global semiconductor market capitalization in recent years. Its valuation core is not wafer production capacity but rather dominance over AI computing architecture and hardware-software ecosystems. Similarly listed on U.S. stock exchanges, Taiwan Semiconductor (TSMC), despite being capital-intensive, enjoys a significantly higher long-term valuation premium than traditional manufacturing due to its irreplaceability in advanced processes.

Another notable example is Broadcom, whose rising market cap ranking reflects the market’s re-pricing of the 'stable cash flow + key infrastructure chips' model. By contrast, Intel's relative decline within the top 50 highlights the market reassessment of its long-term competitiveness following the loss of process leadership.

In the context of Hong Kong and Chinese investment, while SMIC's market capitalization still lags behind global leaders, the key to whether it can remain in the upper observation range for an extended period does not lie in short-term profitability but rather in whether it can establish sustainable technology nodes and secure its role in the industry with policy support. These examples demonstrate that the structure of market capitalization has itself become a crucial indicator for interpreting technological evolution.

When taking a longer-term view, semiconductor companies that consistently rank among the top in market capitalization often occupy irreplaceable positions within the supply chain, rather than merely benefiting from a single wave of demand. Although the market capitalizations of Qualcomm and Texas Instruments may not experience dramatic fluctuations, their long-term stability stems from the depth of their technological platforms and customer structures.

This also highlights that the long-cycle valuation approach for semiconductor stocks does not rely on predicting when the next upturn will occur but instead focuses on determining whether a company is positioned at a critical node in technological evolution over the next decade. As artificial intelligence, cloud computing, and automation become the underlying infrastructure of the global economy, leading semiconductor companies' valuations naturally align more closely with tech platforms rather than traditional cyclical stocks. Analyzing market capitalization structures may reveal more about the semiconductor industry's true long-term value than any single financial metric.