Limited Returns on US and China's AI Investments! Are Taiwan, South Korea, and Japan the Biggest Winners?

In recent years, the United States and China have competed fiercely to invest astronomical amounts in artificial intelligence (AI) infrastructure, a trend now clearly reflected in economic data worldwide—yet the extent of benefits varies significantly. Although the US leads in total investment volume, its actual boost to economic growth has been relatively limited. In contrast, semiconductor-export powerhouses such as South Korea, Taiwan, and Japan have enjoyed far greater-than-expected growth gains in the short term.

According to statistics from Stanford University’s Institute for Human-Centered Artificial Intelligence (HAI), between 2013 and 2025, private-sector investment in AI-related fields in the US totaled $757.2 billion, ranking first globally. Notably, in 2025 alone, $285.8 billion was invested—accounting for nearly 40% of the total.

China follows with $131.8 billion, less than two-tenths of the US figure. However, if official government fund investments are included, the actual amount could be substantially higher.

Despite this flood of capital, the US economy’s tangible gains from AI remain underwhelming. Morgan Stanley estimates that spending on data centers and other AI-related infrastructure could have boosted US GDP growth by 1.37 percentage points. However, because the US heavily relies on imports for advanced chips and communication equipment needed for data centers, once the negative drag from increased imports is subtracted, the net contribution of AI investment to US economic growth may be as low as 0.16 percentage points.

Notably, as data center projects roll out one after another, demand for cutting-edge chips and servers exceeds supply, pushing prices steadily upward. This, in turn, exerts inflationary pressure on products and services incorporating these components.

Kevin Warsh, Chair of the US Federal Reserve, has repeatedly emphasized that while AI adoption will long-term enhance productivity and lower prices, in the current environment, the inflationary effect from surging demand is clearly more pronounced.

Christopher Waller, a Fed governor, even warned that if AI investment continues to surge uncontrollably, it could become a new driver of inflation.

In contrast, China presents a different picture. The country has integrated AI and high-tech solutions into factories and logistics systems, boosting both industrial output and export competitiveness.

Analysis by the Institute of International Finance (IIF) shows that digital industries—including data center construction and communications equipment production—now account for over 10% of China’s GDP, contributing nearly one percentage point to overall economic growth annually in recent years.

However, this momentum is insufficient to fill the domestic demand gap left by a sluggish property market, and the virtuous cycle of “AI-driven productivity gains leading to wage growth” has yet to materialize in China.

Like the US, China remains heavily dependent on imported high-end chips for its AI investments, resulting in a persistent trade deficit in semiconductors.

Crucially, the real beneficiaries appear to be economies with strengths in semiconductor exports—South Korea, Taiwan, and others.

The latest edition of the International Monetary Fund’s (IMF) World Economic Outlook reveals that South Korea, Taiwan, Thailand, and Malaysia posted average annualized GDP growth rates in Q1 of this year that were 4.4 percentage points higher than the IMF’s original forecasts.

South Korea stands out most prominently, with an annualized quarterly growth rate of 7.5%—four times the IMF’s projected figure. Taiwan’s export value in the first half of this year surged 47% year-on-year, with about 80% coming from information technology equipment such as servers and semiconductors, demonstrating the powerful pull effect of AI demand.

Japan is also expected to benefit indirectly through exports of semiconductor manufacturing equipment and electronic components. Nomura Securities economist Issei Ito estimates that if US tech giants like Amazon (AMZN-US) and Alphabet (GOOGL-US) proceed with their planned 2026 AI investments, Japan’s related exports could grow by 17%.

There is widespread hope that this AI boom could recreate the 1990s US-style “non-inflationary economic growth.” Long-term, there remains strong potential for AI to significantly lift the US economy.

Yet markets simultaneously worry that the actual improvements in corporate profits and productivity from AI may fall short of optimistic expectations.

Indeed, since July, previously overheated speculative capital in AI-related sectors has shown signs of reversal, with AI and semiconductor stocks falling sharply across global equity markets. Doubts over whether “AI investment is overheated” are gradually surfacing.

The IMF has also issued a warning: should market expectations for AI dim, not only the US and China but also semiconductor supply chain hubs like South Korea could face risks of stock market corrections and weakened consumer spending due to falling share prices.