World Bank raises East Asia growth forecast on AI exports, warns of tech spending reversal risk

The World Bank has raised its growth forecast for the East Asia and Pacific region on the back of artificial intelligence-related exports, while warning that its reliance on the AI boom leaves it vulnerable to a potential global tech spending reversal.

The region includes 23 economies, including China, Vietnam, Indonesia, Malaysia and Thailand.

The EAP economy is expected to expand 4.5% this year, 0.3 percentage point more than the bank projected in April, according to its latest report released Tuesday. Growth is forecast to ease to 4.4% in 2027 and 4.3% in 2028. Vietnam received the biggest forecast upgrade among major economies of the region, up 1.1 percentage point to 7.4%.

The region's strength, however, is highly dependent on AI-related manufacturing and exports. Trade growth, excluding AI-related goods, has been "weak or negative," the bank said. Those products accounted for more than half of the export growth in most of the region's economies and more than 70% in Malaysia, the Philippines, Thailand and Vietnam.

China, Indonesia, Malaysia, the Philippines, Thailand and Vietnam shipped $1.4 trillion of AI-related goods in the 12 months through April, according to the report.

Official data showed that South Korea's exports grew 83.5% in September to a record $120.9 billion, with chips making up half of those shipments. Reflecting the dominance of semiconductors in the country's market, the World Bank highlighted that just two chipmakers — Samsung and SK Hynix — accounted for 43% of the benchmark Kospi index's value as of end-April.

The AI risk is on the spending side. AI-related capital expenditure has reached about 6% of U.S. GDP, similar to the 2000 peak in information-technology investment, and the current cycle "has risen faster than either previous cycle and is still gaining speed," the bank said.

The Bank for International Settlements in its annual economic report in June had warned that the boom's scale and pace bears resemblance to the dot-com frenzy of the 1990s and other "manias."

The financing driving the boom is also less transparent. Of the $2.9 trillion in AI capex planned for 2025-2028, $800 billion is expected to come from private credit, the bank said, where AI-related lending rose to 34% of activity in 2025 from an 18% average over the prior five years. Private credit portfolios have experienced markdowns, outflows and defaults this year.

Private credit markets are "less visible, and have not been tested by a severe downturn," the bank said.

That said, the AI boom supported by abundant liquidity could slow due to the latest tightening of financial conditions as major central banks raise rates for the first time since 2023, according to the report. The U.S. Federal Reserve raised rates last month, its first increase in more than three years, and signaled one more to hike this year.

A correction may not necessarily mean a bust for the AI supercycle, but that investment "had run ahead of realized demand," the organization said.

A slowdown by 1 percentage point in U.S. growth cuts other emerging-market growth by an estimated 0.6 percentage point, with the hit to investment about twice as large, the bank said. "A slowdown concentrated in AI would be material for East Asia because of the region's prominence in the AI supply chain."

Bank funding is the broadest exposure. Foreign-currency-denominated liabilities of banks appeared significant in some countries — 29.2% of GDP in Malaysia, 20.7% in the Philippines.

Taiwan's statistics bureau recently raised its 2026 growth forecast to 11% from 9.6% on AI demand, while warning in June that "if the high-tech sector faces headwinds, the negative impact on the local economy could be bigger than expected."