Taiwan is on the list! The United States is spending money to develop AI, but the biggest winners are overseas. ING named "these two countries"

Recently, the market has been heatedly discussing the contribution of technology investment to U.S. economic growth, but ING pointed out that U.S. trade data usually does not attract much market attention, but the latest figures are quite shocking, because the overall U.S. trade deficit reached US$105 billion (approximately NT$3.4 trillion) in September, of which technology products alone caused a deficit of US$64 billion (approximately NT$2.9 trillion), accounting for about 61% of the overall deficit, nearly two-thirds. It is worth noting that although U.S. companies invest heavily in artificial intelligence (AI) and data centers, a large amount of related equipment relies on imports, which not only offsets part of the investment's contribution to U.S. economic growth, but also benefits technology product exporters such as Taiwan and South Korea.

According to ING analysis, relevant data shows that as data centers continue to be constructed across the United States, the import volume of technology products is not only showing exponential growth, but the growth curve is even almost vertically upward. Less than two years ago, the United States’ monthly imports of such technology products were approximately US$20 billion (approximately NT$654 billion), but by September this year it had increased significantly to US$87 billion (approximately NT$2.8 trillion), indicating that the demand for equipment driven by AI infrastructure is rapidly expanding.

ING pointed out that the outside world continues to discuss how much the investment boom in technology, AI and data centers has contributed to U.S. economic growth, but in fact, a considerable part of the added value comes from overseas. As the United States purchases large amounts of imported equipment, funds flow to South Korea, Taiwan and other places, which has a negative impact on U.S. economic growth. Among them, U.S. imports from South Korea increased by 35%, and imports from Taiwan increased by an even greater 60%.

Judging from the contribution to the U.S. gross domestic product (GDP), technology capital expenditures and AI Token expenditures, including chips used in data centers, do support economic growth; the concrete, steel and other engineering investments required for data center construction also bring positive contributions.

However, due to the huge scale of the trade deficit in technology products, the current net export project is causing a significant drag on the overall economic growth of the United States. Among them, a considerable number of imported technology products have actually been converted into inventory and are expected to be put into use in the next few quarters. In other words, although AI-related capital expenditures are increasing rapidly, a large amount of equipment must be purchased overseas, causing the growth contribution of the U.S. economy from this wave of investment to be offset by the surge in imports of technology products.

ING emphasized that the large-scale expansion of U.S. technology infrastructure has become an important growth driver for technology-oriented exporters, especially Taiwan and South Korea. It is worth noting that only three years ago, the United States was a net exporter of technology products. Now, due to the surge in imports of technology products, there has been a significant change in the trade structure.

Regarding the U.S. Democratic Party’s intention to increase data center supervision, ING emphasized that the devil is in the details. The analysis pointed out that the most likely measures are target-specific regulations, many of which are already taking shape, including levying data center-related taxes and fees, reducing tax incentives, formulating environmental standards, and requiring developers to bear corresponding cost increases.

ING believes that outright bans are unlikely to become mainstream; even if implemented, they are more likely to be temporary to allow states time to develop a data center policy framework. The article bluntly stated that stricter AI regulation is unlikely to fundamentally change the growth trajectory of the industry, because AI is still very important to the U.S. economy and U.S.-China technology competition.

Data centers that have been licensed or are under construction may be more valuable. New projects are restricted and existing facilities may benefit. As for comprehensive federal legislation, this would likely be subject to a presidential veto, so local regulatory measures may be easier to implement than a national ban.