A-GSEMI: Thailand's First Global Semiconductor Fund

As the world enters a new technological era, semiconductors remain the foundational heroes driving continuous advancement. Over the past three decades, from desktop computers to smartphones, worldwide chip sales grew steadily from roughly $50 billion in 1990 to $440 billion in 2020.

Today, the explosive wave of Generative AI is reshaping growth velocity. Global chip sales reached approximately $796 billion in 2025 and are projected to leap to $3.38 trillion by 2030, marking a remarkable 4.3-fold expansion in just five years—an acceleration unprecedented in history.

Every AI query sent by users requires processing inside massive Data Centers packed with millions of chips. As user adoption surges, these server hubs must expand correspondingly. However, data from JLL indicates a severe global shortage, projecting that demand will reach 220 gigawatts (GW) by 2030 against an available capacity of only 137 GW.

To bridge this infrastructure gap, major cloud service providers including Amazon, Microsoft, Alphabet, Meta, and Oracle are aggressively ramping up capital expenditures. In 2026, hyperscaler investments are projected to soar to $806 billion—a 96% increase year-over-year—before climbing to $1.1 trillion in 2027. Concurrently, the processing chip share within Data Center budgets is expected to rise from 43% in 2022 to 60% by 2027.

Stock photo for illustration only, not from the actual event

Building a single AI chip, however, requires a complex global journey spanning multiple countries and specialists. The process begins with chip design in the United States by industry leaders like Nvidia, AMD, and Broadcom. Manufacturing machinery is supplied by US firms, Japanese players, and ASML from the Netherlands—the sole global producer of extreme ultraviolet (EUV) lithography systems. Raw wafer materials originate from Taiwan, China, South Korea, and Japan, before fabrication by foundries such as TSMC and final packaging by ASE and Teradyne.