Taiwan and South Korea: Extreme Dependence on AI Chips Worries Analysts

The unstoppable rise of AI chips

In January 2026, Taiwan recorded record exports of $65.77 billion, a year-on-year increase of 69.9%, the highest monthly figure in its history. South Korea, meanwhile, sees nearly 60% of its total revenue coming from AI-related products. These figures, reported by The New York Times and cited by WWWhat's new, reflect an unprecedented economic concentration in modern history. To put it in context, in 2023 Taiwan's semiconductor exports accounted for about 40% of its total revenue; in just three years, that share has doubled. South Korea, which depended on chips for 35% in 2020, has seen AI push that figure to 60%. This explosive growth is not an isolated peak but the underlying trend of an entire region that has become the world's AI factory.

TSMC and Samsung: the giants behind the dependence

Taiwan Semiconductor Manufacturing Company (TSMC) produces over 90% of the world's most advanced semiconductors, powering data centers for Nvidia, AMD, Apple, and other major tech companies. In South Korea, Samsung and SK Hynix dominate high-bandwidth memory (HBM), essential for AI accelerators. Taiwan's Taiex index and South Korea's Kospi hit record highs in 2026, moving as proxies for US hyperscalers' AI spending. For instance, the Taiex rose 28% in 2025 and another 15% in the first quarter of 2026, almost perfectly correlated with data center investments by companies like Microsoft, Amazon, and Google. According to SEMI data, the global chip equipment industry expects sales of $126 billion in 2026, with China, Taiwan, and South Korea as key markets. TSMC alone accounts for over 25% of Taiwan's GDP when considering its supply chain, according to estimates from Taiwan's Central Bank.

The risk of concentration

Analysts at Goldman Sachs, such as Tim Moe, warn that this dependence creates a critical vulnerability. If demand for AI chips slows, even for a quarter, both economies would suffer a severe impact. Moe notes that “Taiwan exposes more than 80% of its revenue streams to AI-related products, while South Korea is around 60%.” This concentration amplifies the risk of external shocks, such as geopolitical tensions (e.g., an escalation in the Taiwan Strait) or changes in US technology policies, which could impose restrictions on advanced chip exports. In 2023, Washington already limited the sale of AI semiconductors to China, temporarily affecting Nvidia's stock and, by extension, TSMC. A similar but broader scenario could have devastating consequences. Moreover, dependence on a single buyer—the United States—exacerbates the risk: according to WTO data, 40% of Taiwan's chip exports go to the US, and 30% of South Korea's.

Implications for the future

The current situation echoes the dot-com bubble, but with a key difference: AI demand is backed by massive investments from big tech. In 2025, hyperscalers invested over $200 billion in AI infrastructure, according to Synergy Research Group, and that figure is expected to grow 30% in 2026. However, the lack of economic diversification in Taiwan and South Korea is a warning sign. Historically, excessive dependence on one sector has led to crises: for example, the dot-com bubble in 2000 dragged down Taiwan's economy, which then relied on computer hardware exports for 30%. Today, the dependence is greater and more concentrated. For investors, geographic and sectoral diversification is more important than ever. For readers, understanding this dynamic is essential to anticipate potential global economic crises. As Tim Moe concludes, “if demand for AI chips slows, even for a quarter, the impact would be severe.” The question is not whether it will happen, but when and how deep the adjustment will be.