AI data center spending lifts exports, raises recession risk

By Amara Osei

As much as 80% of the recent increase in global exports has been concentrated in AI-related goods, according to Capital Economics, the British economic research firm cited in a Wall Street Journal report published on the 11th.

The firm projects global trade growth this year at its strongest rate since the early 1990s, excluding rebounds that followed the COVID-19 pandemic and the global financial crisis. The forecast rests on expanding U.S. data center investment, which has lifted exports of semiconductors and server-related equipment from South Korea, Taiwan and China.

Those three economies have recorded the largest export gains and are seen as the most exposed to an AI spending reversal. South Korea's total exports jumped 93.7% in October from a year earlier, while semiconductor exports surged 259.4%. Chinese exports rose 35%.

How much of the trade boom comes from AI?

Capital Economics estimates that AI-related goods account for up to 80% of the recent increase in global exports. That concentration links trade performance closely to one spending cycle: investment by large U.S. technology companies in data centers.

If those investments contract, Asian economies dependent on the associated exports could slow sharply. The exposure reaches beyond AI companies to national export totals and the wider global growth cycle.

One spending cycle now accounts for as much as 80% of the recent export increase.

Why could an AI slowdown trigger recession?

Capital Economics warned that a cooling in the AI spending boom could tip the world economy into recession. The transmission runs from U.S. capital spending to Asian factory orders, export revenue and regional growth.

“A sharp drop in AI-related share prices, combined with other negative shocks, could lead to a global recession,” said Gaurav Ganguly, head of international economics at Moody's Analytics.

Frederic Neumann, chief Asia economist at HSBC, said the global economy would struggle to avoid a downturn if the AI boom were to unravel.

Both economists identify falling asset prices or reduced investment as channels through which a downturn could spread. A pullback in data center spending would also remove demand for the semiconductors and server equipment at the center of the current trade expansion.

A reversal in U.S. data center budgets could travel quickly from Asian export orders to global growth.

What does the dot-com precedent show?

The report draws a parallel with the collapse of the dot-com bubble. Oxford Economics found that U.S. imports of information technology equipment as a share of gross domestic product peaked in mid-2000 and then fell 55% over two years.

The pullback in U.S. investment also hit Asian exporters, and global economic growth slowed to 2% in 2001. The sequence links a technology investment retreat, weaker Asian export performance and slower worldwide growth.

The dot-com precedent does not establish that AI spending will produce the same result. It does show that a technology-led investment cycle can transmit losses across borders when exporters depend heavily on U.S. demand.

The dot-com episode shows how a technology spending retreat can spread across borders and slow world growth.

Which economies carry the most exposure?

South Korea, Taiwan and China sit at the center of the trade gains because their semiconductor and server-related exports have benefited from U.S. data center investment. A cut in that investment could reduce demand for those products and slow export growth across all three economies.

The vulnerability extends beyond one product market. U.S. data center investment supports demand for semiconductors and server-related equipment from each of the three exporters, concentrating their exposure in the same corporate spending decisions.

For South Korea, the concentration appears in the October semiconductor figures. Total exports rose 93.7% year on year, and semiconductor exports rose 259.4%. Chinese exports increased 35%, while Taiwan also ranks among the largest beneficiaries.

For exporters and policymakers, the relevant measure extends beyond AI investment alone. It includes the share of export growth and GDP tied to a spending cycle that can reverse. Capital Economics puts the export concentration at as much as 80%.

Current trade data supports growth, but the risk lies in its dependence on a single investment cycle.